->

-Iz
(,/)0 ::AlrJ
{./lIz
go r-r-1

...... ,
f=~IZ "r ':.•
:.Y _" . :.:'V': _ .'. :.:'V': _•

o r l....,
-:"Il .....
\.~
,,:....·,:.~"r ~ "f ':.~ 'j
.~

:,
\ •... ..~/.:.\..~
I..... -...J ~,
,""
" ,,~...,
"" ":, ."" ":, ."" ,.
~.... ,,~.'" '\.i
.... ~ I · ow I....· ow I....·'
.".~~.l.
.".
,. ..•. . ' . ..... ... ' ....".. ~"l
.,. ~"'- ". ' ....,... ~"1
...., ' ....,... ".
'l\ r. :. . r.
.'.. :::0
..,.,
•• • ., • "1
'."',
,,~.'" "'"
,,~.,
••' ,r••,,\o\:\•.•• "~~\."" "~~\.""
t ...·"V I....· ow I....· ow I..
••, -...J I....· -...J "

... ~l·~~1 OIr-r-1 '(., 'R:"l ~"1 ~..

·,:~_·'r. ',:~'\ ,\,
••••• - '
,. :, r. '.., ,. ":, ".. :1_
•..
•..... ~~\ , ..~~\ ~ •..... ..
i.•..•..
I....·
."\:.. ~~.~~
~-
~~\.•.. ,
-...J
.
':."-,':,,,-,''''''
i r. :.,.,. :,'"
"~A..
i....· "'"l
~- . \:V
r 10 ~.,- ~
\"-. :
i....'·
, ,
...., I....· ow I....' ow I....·
..
- -'.,. ,. ".., r. ~ .. -"
'1
"~~\."" "~~\."" ..~~ "~A:\.'" "~A:\..'" .. ~~\. •..•
I....' ow I....· ow I.

3 0 lrJ
••, -...J i...., i....·....:.
-...J •••• "V
.. .. ~~ "~'l~'l~'
~~~\. •..,:.,.,.
~~
-.-.. • r_
:.1-·.. • :...... .o..
~_
. .""
':~_"'r ':,~."'f. ':.~j
. :.,. r.
:\.•..•..

, "1. ~'l
.·~·r···
'. " ' ,
:.,.
"~"\"" "~A
I....· -...J i....·
..,. ". :.,".- ~'l
• '."',
-..J

"
I....· ....
". ". :',.... 'R'
..,. ":..
.r•• " ' \ • '. u:
{./l r-r-1
~
":,
~

~f. ~l. ~:.(.

~,
,.:\ ."" ~"
."" ,,~."
! I ' ow I · ow I....·

':.~-", .:.~ . ', .

.~~\ ••••.'••~"-\ ••••.'•• ~~\. ••••J
'.
,,~.'" ,,~.'"
••, -...J I....· ~ i....·,,~." ,,~
-..J I....· .... .."V , ..... ow I....' ow I.
 SCIENTIFIC
-KNOWLEDGE-
ITS SOCIAL PROBLEMS
 Jerome R.
RAVETZ
SCIENTIFIC
-K NOW LED 0 E-
~s SOCIAL PROBLEMS
WITH A NEW INTRODUCTION
BY THE AUTHOR

•
Transaction Publishers
New Brunswick (U.S.A.) and London (U.K.)
New material this edition copyright C 1996 by Transaction Publishers,
New Bnmswick, New Jersey 08903. Originally published in 1971 by
Oxford University Press.
All rights reserved under International and Pan-American Copyright Con-
ventions. No part of this book may be reproduced or transmitted in any
form or by any means, electronic or mechanical, including photocopy,
recording, or any information storage and retrieval system, without prior
permission in writing from the publisher. All inquiries should be ad-
dressed to Transaction Publishers, Rutgers-The State University, New
Bnmswick, New Jersey 08903.
This book is printed on acid-free paper that meets the American Na-
tional Standard for Permanence of Paper for Printed Library Materials.
Library of Congress Catalog Number: 95-32177
ISB~: I-S~8S1-2
Printed in the United States of America
Library of Congress Cataloging-in-Publication Data
Ravetz, Jerome R.
Scientific knowledge and its social problems I Jerome R. Ravetz ;
with a new introduction by the author.
p. CID.
Originally published: Oxford : Clarendon Press, 1971.9
Includes bibliographical references and index.
ISB~ I-S~851-2 (pbk. : alk. paper)
1. Science-Philosophy. 2. Science-Social aspects. I. Title.
QI75.R3 1995
SOI-dc20 95-32177
CIP
 To the memory of my father
GUS RAVETZ
who taught me to think for myself
 CONTENTS

Introduction to the Transaction Edition ix

Preface xxv
Introduction I

PART I. 1HE VARIETIES OF SCIENTIFIC EXPERIENCE 7

I. 'What is Science ?' II

2. Social Problems ofIndustrialized Science 31

PART II. 1HE ACHIEVEMENT OF SCIENTIFIC XNOWLFDGE 69

3. Science as Craftsman's Work 7S
4: Scientific Inquiry: Problem-Solving on Artificial Objects 109
5. Methods 146
6. Facts and their Evolution 181
7. The Special Character ofScientific Knowledge 209

PART III. SOCIAL ASPECTS OF SCIENTIFIC ACI1VITY 241

8. The Protection ofProperty 24S
9. The Management ofNovelty 260
10. Q!tality Control in Science 273
II. Ethics in Scientific Activity 289

PART IV. SCIENCE IN THE MODERN WORlD 31 5

12. Technical Problems 321
13. Practical Problems 339
14. Immature and Ineffective Fields ofInquiry 364
PART V. CONCLUSION: 1HE rtmJRE OF SCIENCE 403
Index of Names 437
Index of Topics 443
General Index 446
 INTRODUCTION TO THE
TRANSACTION EDITION

Where the Book has Kept Its Relevance

THE book was written to help myself and others comprehend
changes in the life of science that were getting under way in the
early 19708, and which are still developing. For science was
then visibly changing from being mainly a small-scale endeavor
of individual scholars to being a large-scale enterprise of teams
of researchers. My big problem was how, under these new con-
ditions, science could preserve the qualities that have made it
such a valuable element of our civilization. And since the
changes are coming in the social organization of science, I had
to see how the social aspects of science condition the way that
scientific knowledge is achieved. All this is still relevant today;
and although in some areas scholars have produced more detail
and new insights, there is very little in the book that feels obso-
lete or out of touch.
I would like to provide some guidance for prospective read-
ers, identifying those parts of the book that will be most useful
now. But it is not too difficult to explore the rest, for there are
summaries at the beginning of every subdivision of the text.
Going through these will give the reader a quick survey of what
is there; then one can decide whether to explore more deeply.
For particular sections, I can still recommend Part I, "The Vari-
eties of Scientific Experience" as a good read on important prob-
lems. There I run through the contrasting, sometimes conflict-
ing "images" of science, and then rehearse some of the social
problems of science that I later explore and explain in depth. I
remind readers that the opposition of an idyllic pre-war "little
science" with a morally compromised "big science" or "indus-
trialized science" is overly simple in historical fact; but it is
still important in the images by which science is defmed and
defines itself. For those who are not yet fully immersed in the
world of postmodemity, who still believe that some things in
x Introduction to the Transaction Edition
this world can have qualities of goodness that need defending,
the nanative of Part I is still alive.
The most successful substantial section of the book, and per-
haps the one which will live the longest, is Part n, "TheAchieve-
ment of Scientific Knowledge." This is based on the idea of
"science as craft work." With "science" interpreted as scholar-
ship or research in the broadest sense, my analysis was not only
well received by reviewers, but has also stood the test of time.
It has its limits, for it applies primarily to "pure" or "basic"
research ofthe sort traditionally conducted in universities. Now
that that sort of research is shrinking in size and importance,
my discussion would need to be modified to suit the new con-
texts of mission-oriented research and what I am now calling
"post-normal science." But I am as certain as ever that my vi-
sion of science as craft work is necessary if we are to get away
from the idea that "scientists discover facts." Knowledge is not
an automatic outcome of research but rather the product of a
lengthy social process, of which the major part occurs after,
even long after, the research is completed. So long as students
are conditioned, by philosophers as well as by their teachers, to
think of scientific infonnation in tenns of "true or false," we
will be severely hampered in our attempts to apply science to
the solution ofthe pressing problems of planetary survival. The
key idea is to substitute "quality" for "truth" in the evaluation
of scientific materials; the social and ethical aspects of science
are then firmly on the agenda, as well as scientific uncertainty.
The next section, "Social Aspects of Scientific Activity," was
not so uniformly appreciated; but it was a pioneering effort to
consider issues like quality control and ethics in research. It
was rather tightly focused on traditional research, but the prob-
lems it discusses are very relevant to science as it is practiced
now. At the time I wrote that section, problems of quality in
science were ignored by philosophers and only infrequently
discussed publicly among scientists. But I had already seen,
from my experience of several fields, that quality control can-
not be taken for granted; research of astonishingly low quality
can be perfonned, published, and rewarded. At the end ofa long
and difficult analysis of quality control in science I came up
with a surprising conclusion. For it turns out that quality con-
trol in science is a sort of bootstrap operation, and depends on
 Introduction to the Transaction Edition xi
the morale and commitment of working scientists, as reinforced
and ultimately sustained by the moral quality of the leadership
ofthe scientific communities. This actually emphasized the prob-
lem that was one of the main motivations for the book; for if
the old idealism of "little science" has lost its social and ideo-
logical foundation, what corresponding source for idealism is
there in the "industrialized science" ofthe present? Without such
an idealism, science would be very vulnerable to corruption,
leading to universal rule by mediocrity or worse.
This question throws up another one that is directly relevant
to our current concerns in the use of science; and although the
book does not provide an answer, it enables us to frame the
problem. When we consider the issues in which science is em-
ployed, as with risks and environmental policy, we find a situa-
tion that is very different from that of the craftsman-scientists
of old, who chose their problems and then investigated them
under the guidance of the criteria of value and adequacy estab-
lished by a communal consensus of their peers and mentors. In
policy-relevant research, that haven is no more; now we have,
typically, "facts are uncertain, values in dispute, stakes high,
and decisions urgent." The fields of science that are employed
tend to be low in prestige and in strength, and are frequently
what I call "immature" or "ineffective." Now, here is the prob-
lem: if quality control in traditional research depends on such a
special community, sheltered from the harsher realities of the
world of affairs, and led by persons of ability and commitment,
then how is there to be any effective quality control where ev-
erything is partisan, contested, and conflicted? This problem is
quite serious, for if quality control of science fails and is seen to
fail in these policy debates, then by default it will be brote p0-
litical power that decides. As I was writing the concluding sec-
tion of the book, I did see the problem, and I even identified it
in a way that one reviewer used against me very cruelly (see
below). But as I then still had so little experience of the "critical
science" I was advocating, I did not appreciate the enormity of
the problem after until the book had been published and I was
thinking of a sequel.
In that way, the book failed to solve the deep problem that
had originally motivated its writing: how can the essential quali-
ties of science be maintained under conditions so different from
xii Introtluction to the 7Tansaction Edition
those of its foundation and maturing? My consolation is that
this is not the first such failure in the history of philosophy! But
that failure, when I became aware of it, was one of the things
that inhibited my attempts to write a sequel to Scientific Knowl-
edge. I felt that I could get away with one somewhat narcissis-
tic exploratory venture, but the next one should deliver what it
promised. All this is very relevant to the present time, for no
one else has yet gone deeply into that new problem of quality
control in policy-relevant science, nor even gone beyond my
philosophical analysis ofquality control in research. So the chal-
lenge is there; I am working up ideas on it, and I would wel-
come a dialogue.
Not long after the book appeared, I wished that someone had
advised me to let humanity wait for my message a few months
more, while I let it rest on a shelf; then, somewhat distanced
and refreshed, I could have gone over the text with a blue pen-
cil, and perhaps a pair of scissors, and pruned it to good effect.
There is much material in the latter half which helped me to
satisfy myself on some problems, but which is either too dense
or just not important enough (or both), and which presents an
obstacle to the reader. A whole section that I considered quite
important, "Science in the Modem World," could have been
drastically revised, particularly since I could not speak from
close personal knowledge about the phenomena I was discuss-
ing. In spite of those weaknesses, I believe that my essential
points are still valid; perhaps as the common sense of science
evolves, arguments which then seemed abstruse will become
obvious. For example, the tenn junk-science is coming into
currency in the United States; this means that all the phenom-
ena of low-quality research, and then (as in this latter section)
immature and ineffective disciplines, will have a resonance. So
this section too can still carry a message, perhaps even more
now than at the time of its writing.
The nanative picks up again in the concluding section, and
particularly in the unashamedly speculative last pages. This re-
ally goes at a gallop (reflecting the circumstances of its compo-
sition), ftrst with a very self-critical account of "critical sci-
ence" and finally with a vaguely prophetic message as a substi-
tute for a call to action. (This was noticed with disapproval by
some fonner comrades in the radical science movement). I think
 Introduction to the Transaction Edition xiii
both sections are still relevant and important today, though I
believe that with the development of "post-normal science" I
could be more specific in my recommendations. I have often
advised students that when confronted with a thick book they
should start with the first and last chapters and the introduc-
tion; and that would not be bad advice for this one.
The Parts That I Like Most
There is no guarantee that prospective readers will agree with
my evaluations, but there are some things that I believe are
worthy of special notice. First, I do like the title, which was
very bold in its time. People were just getting around to admit-
ting that scientific knowledge could cause problems (as "the
bomb"), and that scientific research is in some respects a social
activity with some ethical problems; but to say that knowledge
could have social problems would seem illogical. I don't recall
defending the title explicitly in the book; but it is all about sci-
entific knowledge as the outcome of a social activity, and so
with only a slight stretch one could imagine how the knowl-
edge itself could have problems of a social character, such as
the failure of quality-eontrol mentioned above.
Another shocking thing about the book, as a work purporting
to be scholarly, was the footnotes: their relation to the main
text, and their contents. Some reviewers observed that the foot-
notes seemed to be appended to a freely flowing text, almost as
if students had been sent out to collect examples for an argu-
ment that was already completed. This overestimates the qual-
ity of student labor; but it is fair to say that while writing the
text I did make notes of ideas for footnotes to be written and
appended later. This was a consistent policy; I decided that the
examples could only illustrate my argument, not prove it. To
incorporate and argue the examples inside the main text would
make it impossibly long and complicated. Also, since the text
was covering a wide range of topics in an original fashion, it
was inevitable that the contents of the footnotes would be un-
conventional. I confess that I had a lot of fun with the foot-
notes; my favorites include the anonymous tram driver in At-
lantic City with his valuable aphorism, the idea of a sequel to
An Enemy of the People, and Lincoln Steffens' explanation of
political corruption. I do recall an academic colleague telling
xiv Introduction to the 7Tansaction Edition
me that the inclusion ofjokes in my footnotes meant that mine
was not a serious work.
There were some real nuggets of insight there also, some of
them expressed as lists of categories or distinctions. Thus, I of-
fered a sort of"four horsemen ofthe scientific apocalypse" in the
form of shoddy science, entrepreneurial science, reckless science,
and dirty science; and they are all involved with runaway tech-
nology. I gave a nice definition of entrepreneurial science, that
which occurs when an series of scientifically adequate research
projects enables a series of financially successful grants, rather
than the successful projects depending on adequate grants. The
other terms need no explanati~n.In the discussion of science in
the modem world, I analyzed three sorts ofproblem, "scientific,"
"technical," and "practical," having to do with knowledge, func-
tions, and purposes respectively. The latter two categories are
different sorts of "final causes"; and much of the enthusiasm for
applying science to human betterment has involved ignoring the
deep difference between functions and purposes.
A very fmitful idea that I used in connection with "methods"
is "pitfall." This is not my own; my Ph.D. supervisor in math-
ematics had the aphorism, "In analysis the pitfalls are every-
where dense." Having done research in that field, I well knew
what he meant. How many times did I write up the "final" draft
of a paper, only to discover yet again that a statement described
by me as "obvious" was not only obscure but also wrong! I was
gratified to discover later that the concept had been taken up by
people in the field of applied systems analysis to investigate
their own difficulties; and that idea, combined with the craft
character of research, has had a useful life in the theoretical
reflections in operations research. Less glorious, perhaps, is the
history of the compound term complex and subtle. I needed to
repeat this again and again, as antidote to the traditional propa-
ganda, of scientists and philosophers, that science is simple,
either in its results or in its method. I paid the price for this
when a graduate student in Leeds went through the text of the
book looking for occurrences of the term, and reported the re-
sults ofhis count in an effectively public fashion. Some review-
ers also took exception to "deep".
There is one section of which I am really rather proud, as I
consider it to be real philosophy and not just a generalized de-
 Introduction to the Transaction Edition xv
scription of how things are. Interestingly, this has had hardly
any comment, in spite of coming at the end of the most acces-
sible part and growing out of it. Indeed, it provides the answer
to the problem that motivated that part of the book: how the
intensely personal and subjective activity of research can give
rise to the impersonal and objective knowledge that is science.
Perhaps the answer turned out to be too paradoxical for people
to hold onto then. But I feel that it deserves a new lease of life,
as a possible basis for an answer for the nihilistic tendencies in
deconstruction and postmodernity.
It starts with an analysis of "fact," which is some way down-
stream from a completed and published research report. (I will
not go into the earlier phases of the work, since they are in the
section that seems to be easier to understand). Most such items
die at that stage, as they are never used or cited; whatever their
internal quality, they were just insufficiently interesting to sur-
vive through use. (If the general public were aware of the size of
this wastage, there might be more questions about quality con-
trol in publicly supported science, even extending through the
phase of journal refereeing and publication). Those results that
are used are then tested through attempted repetition; and then,
more interesting and significant, how "invariant" they are when
methods and concepts change. H a result is still in use, it can be
considered as a "fact," and it starts another path of evolution to-
ward various functions in teaching and in tool use. In each of
these it becomes "standardized," a sort of robust tool, and the
complications of its original creation become irrelevant and are
stripped away and forgotten. Students then learn a simplified ver-
sion of the fact, presented as if it were an atom of truth. Some
start on the path to phil~phy by discovering that different courses
can teach radically different versions of the same "fact," or even
that different editions of the same textbook can silently move
from one unquestionable version to another.
At this point the real paradoxes begin. First, the process of
standardization necessarily suppresses anything that might be
confusing; and so the presence of obscurities at the foundations
of theoretical science is a taboo subject, among teachers and
publicists. (Some philosophers know about them, but even they
tend to see them as soluble problems rather than as inherent
contradictions of theoretical knowledge). Students who ask too
xvi Introtluction to the 7Tansaction Edition
many awkward questions (as in differential calculus, are we
dividing zero by zero, and if not, what?) may be given the (false)
promise that it will all be explained in the next year's course;
or, failing that, are handled by "proof by intimidation." This
feature, characteristic of most natural science but imitated in
good faith by some social sciences, forces teachers to become
as dogmatic as theologians of old, and produces brainwashed
technicians out ofstudents. The eminent philosopher T. S. Kuhn
also realized this; but contented himself with expressing the
insight in ironic phrases in his great book. (On this, see my
essay "Ideological Commitments in the Philosophy of Science,"
in The Merger ofKnowledge with Power).
Finally, I argue that real scientific knowledge is the outcome
of all these proces.vs of differentiation and standardization, and
establishment of separate pathways back into experience. For a
suggestive metaphor, the elements of scientific knowledge are
not so much building stones laid at the foundation of edifices of
facts, but are more like mangrove trees, interpenetrating and
rerooting themselves as their means of survival and growth.
Such a process of the ceaseless creation and re-creation of sci-
entific knowledge can help us in understanding our current pre-
dicaments. First, the analysis is not restricted to natural sci-
ence. Any learned discipline goes through the same sort of evo-
lution; those that are more historically oriented can be very self-
aware of their own history. Their students can then be tested on
their critical awareness, and not merely on their ability to jump
through conceptual hoops. We can even extend the picture to
great artistic productions that always need to be reinterpreted
afresh, notably drama. When we see a production ofa play more
than two thousand years old and find that it speaks to us, there
is genuine knowledge, not of the rigorous scientific sort, but
very real nonetheless.
I do believe that this enriched historical vision of scientific
knowledge can make a great contribution, not merely to "human-
izing" education in science;-but also in enabling the sharing of
perspectives between scientists and others that will be essential
if the total problems of the environment are to be addressed and
solved. If we are to achieve true communication across disciplin-
ary boundaries, then scientists will need a training that prepares
them to see their field in its contexts, philosophical, historical,
 Introduction to the Transaction Edition xvii
social, and ethical. Fringe courses help in their way, but they will
remain on the fringe until scientists can see themselves, and their
students, as part of a great social endeavor.

What is Missing from the Book

What is missing refers, of course, to things that should be
there and are not. These are of two sorts: those that should be
there for a reader now; and those that might have been expected
to appear then, and did not.
Both lists have a common element: there is not much criti-
cism of the social problems of science as it is practiced. This
might seem to be a serious lack; but for my purposes then they
were not significant. I knew all about sexism and racism in the
communities of science, and I gave each of them a mention.
For women, there is what I call the droit du Professeur, or the
"right of the first publication," analogous to the old droit du
Seigneur celebrated in Mozart's Marriage ofFigaro. And I re-
mark on the traditional exclusion of Jews and blacks from sci-
ence. But these remarks were marginal; clearly, for my analy-
sis, such elements of prejudice and exploitation were acciden-
tal to the life of science, in that they did not make a significant
difference to its work and results. In one sense they were inevi-
table; for even if its members had wanted to, science could not
insulate itself from the general society of its time and adopt
practices that would be seen as aberrant and threatening. Had I
devoted more time to the social and human sciences, I might
have given more weight to these defonnations; but the materi-
als for that history were only beginning to emerge when I was
writing the book.
I confess that in spite of my awareness of these defects, there
is a certain implicit complacency about the elitism of science.
Perhaps this is because I had decided that anti-elitism in sci-
ence still has a long way to go historically before it can become
effective. Earlier attempts at "science for the people" had either
been abortive or disastrous, as in the case of the Soviet agrono-
mist and charlatan T. D. Lysenko, a story that was still very
much alive then. I also had before me the examples of the trag-
edies of the communist scientists of prewar Britain, whose com-
mitment to social change and refonn of science led to their cap-
ture by the Soviet Union. I also knew about the counter-culture
xviii Introduction to the 7Tansaction Edition
of the 19608, and dealt with it at some length in the conclusion
of the book.
In some ways I am a very practical person; if I feel that a prob-
lem is not yet on the historical agenda, I will not make exhorta-
tions about it. I no longer believe that it is useless to make clarion
calls for the achievement of Utopia; I know they might function
as a vision and inspiration to future generations. It is just that
temperamentally I am unsuited to that sort ofwork. Former com-
rades in the radical science movement were annoyed at the ab-
sence ofengagement to the cause; what I said about "critical sci-
ence" was deliberately sketchy, and also explicitly self-critical.
Others really did try for years afterwards, through the 19708 and
beyond, to form a coherent radical vision ofscience; but I moved
away, without trying to make a political movement of "critical
science." Only recently, as my ideas on "post-normal science"
have developed, can I see a way forward to the democratization
ofscience; and that is only in the policy-related fields.
Another significant absence from the book is a dialogue with
the issues discussed by leading philosophers of science of the
period. The decade of the 19608 was a sort of golden age of the
philosophy of science, for Popper and Kuhn were both active,
along with Lakatos and Feyerabend; and these were all intel-
lects of real excellence. I am sure that this omission was costly
to me in the short run, for the book has been nearly totally ig-
nored by the professional philosophers of science. By their cri-
teria, or within their paradigm, it is not a work in the philoso-
phy of science. It is not that I was unaware of the great debates
of the period; I was on terms of acquaintance or friendship with
all the leading members of that philosophical community. But I
simply could not get excited about their problems. I did not
need to dialogue with, or break away from logical positivism or
its rivals, for I had never taken it seriously myself. My own
internal debate had been with Marxism, but that had been settled
by the time that I was beginning work on the book. It could be
that this omission will now help the book to a longer life, for
there are no obviously anachronistic issues being discussed there.
What Reception It Got and Why
The reviewers of the book fell mainly into two classes. There
were those who knew me, and who knew my concerns and style
 Introduction to the Transaction Edition xix
from my previous writings and activities, such as the Campaign
for Nuclear Disarmament, the Anti-Concorde Project, and the
British Society for Social Responsibility in Science (B.S.S.R.S.).
And then there were the scholars, who were perhaps equally
concerned about the state and future of science. They did not
see the book as Jerry Ravetz's long-awaited opus, but only as
yet another exercise in the philosophy and sociology of sci-
ence. Reading the first group of reviews makes me feel as if I
am at a homecoming. For I see evidence of such appreciation,
trust, and affection, that even now I would like to contact them
(for fortunately they are all still alive and well) and express my
thanks. It is a reminder to me that even though the "critical
science" movement never got a lasting organizational structure,
or even a coherent statement of aims and methods, there was
something there, a comradely network that in some ways still
survives.
The academic reviews provoke a different reaction; I blush
as I read their penetrating criticisms. They all agreed that the
book was interesting, indeed worthwhile, but as a piece ofschol-
arship it was more or less seriously flawed. They could not flow
among the varieties of implicit meanings of "science," as my
friends could, so they were confused by the ambiguities in my
use of the term. Nor would they remember the caveats at the
beginning of every section that were intended to cover all the
grand historical generalizations contained therein. They found
an annoying contrast between my praise of careful scholarship,
and the amateur indexing and the unorthodox footnotes. Such
things may seem small, but for someone becoming initated with
a long and difficult book, they provide clues for a negative judg-
ment. One reviewer seems to have tried to work out what the
section titled "Science in the Modem World" was all about, and
decided that it was a veiled attack on social science. He then
read the whole book that way, and was very angry. Rereading
the material, I can see why he made that interpretation; in that
section, I was somewhat out my depth, both with the ideas and
with the evidence. I remember that the distinction between "im-
mature" and "ineffective" sciences came at a very late stage in
the composition of the text; and back then we did not have word
processors with "find and replace" utilities. Also, the sciences
that were used for "practical problems" were mainly the social
xx Introduction to the 7Tansaction Edition
sciences (although I should have said more, much more, about
medicine). Now all the environmental sciences are involved in
"practical problems," and that section could now be quite use-
ful as a basis for discussion of the problems of those sciences.
Best of all was the concluding comment of her review by an
academic historian, who had research experience of some of
the matters on which I was generalizing. She took one of my
warnings about "critical science," and turned it around: "Fi-
nally, talldng about'critical science, 'his Paradise Regained (my
title, not his), he says the most obvious pitfall will be 'an accre-
tion of cranks and congenital rebels whose refonning zeal is
not matched by their scientific skill.' Yes, indeed." There was
only one possible response to that: a postcard simply saying,
"Touche." Afterwards we were best friends.
Other reviewers, while sympathetic, sensed that I had stretched
myself rather thin with my evidence, or noticed the failure of
the book to produce the answer that it had spent so much time
building up to. For the first, an American reviewer remarked
that my prototypical scientist seemed to be like "Dr. Grant
Swinger," the entrepreneurial scientist immortalized by Dan
Greenberg's column in Science. For the latter, an English re-
viewer decided that I am "an incurable romantic, hopelessly
and delightfully in love with science"; that only because I had
specialized in mathematics could I believe that ecology would
become the "critical science"; and finally that I had fudged the
issue of the relations between the radical refonn of science and
that of society at large. The last comment is correct; the previ-
ous two are matters of judgment. It is also worth recording the
experience of a young environmental scientist who enthusiasti-
cally read through nearly the whole book without noticing a
political slant, and who was then devastated by the last chapter!
These criticisms are worth recounting, for they remind me of
the imperfections of the book; and from them I get reminders of
what I should do, and what I should avoid, the next time around.
I also hope that they will be useful to prospective readers; points
of annoyance and frustration will have been anticipated by re-
viewers, and perhaps in this way readers will have been warned.
But when all is said and done, I could not have written such an
ambitious book in a very different way. To do a proper schol-
arly job on such a broad problem would have been the style of
 Introduction to the Transaction Edition xxi
a German professor of a previous age; and I felt that it was
more important to work up my insights in as disciplined and
coherent a fashion as was feasible, rather than to spend decades
on a work that would be proof against criticism of the scholars.
It was a gamble, and, almost all the reviewers agreed, it paid
off; the book was worth reading, and therefore worth writing.
Since Then
The book was at a disadvantage, in its writing and its initial
reception, because I was dealing with problems that were just
then taking shape. Hence, at times I had to be impressionistic
with my evidence, and many readers just could not see what I
was worned about. Some American colleagues in the history of
science tried the book on advanced classes, and from the stu-
dents' responses it was clear that it spoke only to a minority.
But in publication terms it was far from being a failure, going
through two hardback and two paperback editions in the United
Kingdom, as well as a paperback in the United States; it did not
go out of print for more than twenty years. It is still being pho-
tocopied in appreciable quantities, and even a' an increasing
rate. It is quite possible that what seemed obscure or difficult in
the 1970's is now appreciated as relevant and obvious.
I did rather little to promote the book, partly because there
was no defmed disciplinary audience to which I could appeal.
Shortly after its completion, I found myselfengaged on the clos-
est thing to "critical science" that was practical at the time. This
was running "The Council for Science and Society," a small
think tank trying to be a "corporate conscience for the scientific
community" in the United Kingdom. There I had a very fmitful
experience of the strengths and limits of working inside the es-
tablishment, however far over on the margin. As my term there
ended, I found myself on the Genetic Manipulation Advisory
Group, which gave me invaluable experience of the regulatory
game as practiced in the United Kingdom. By this time (later
19708), the B.S.S.R.S. had shrunk to a shadow, and with it the
practical prospects for a politically radical critique of science.
During this time my main scholarly output was essays, usually
invited lectures or papers; I had the good fortune to see these
published in The Merger of Knowledge with Power (London,
Cassell, 1990). The title recalls Francis Bacon's vision of a
xxii Introduction to the 7Tansaction Edition
maniage ofthe two principles; the term merger is ironic, recall-
ing the third great lie of modern civilization: "It will be just the
same after the merger."
Then in the early 19808 I began the work that would even-
tually bring me back to a matured version of "critical science."
This began on the topic of uncertainty in quantitative infor-
mation, which is apparently as nonpolitical as can be. But the
appearance is deceptive, for in debates on environmental ques-
tions, uncertainty is rife, and is manipulated in many ways for
overtly political ends. The book I published with my colleague,
Uncertainty and Quality in Science for Policy (with s. o.
Funtowicz, Reidel, 1990) is mainly technical, but the epilogue
gives it all away. By the time of its publication we were fully
engaged in a more fundamental philosophical exploration, on
the implications of this radical uncertainty for the appropriate
methodology for problem solving in the area of risks and the
environment. Out of this has come the concept of "post-nor-
mal science," of which the core idea is the enhancement of the
peer community responsible for quality assurance of the sci-
entific inputs. (Thus, my original concern with quality con-
trol in science has borne fmit). Instead of being restricted to
discipline-based scientists or official experts, this peer com-
munity should include all the stakeholders in an issue who are
committed to a dialogue. This concept reflects practice in many
places, particularly North America; and as we have articulated
it, people engaged in that practice fmd it illuminating and use-
ful. Hence, for the present, "critical science" finds its realiza-
tion in "post-nonnal science." We have published essays about
it quite a few times; it seems that the most easily accessible
source is one with a rather early version; this is "A New Sci-
entific Methodology for Global Environmental Issues" in R.
Costanza, editor, Ecological Economics (New York: Colum-
bia University Press, 1991).
There are several philosophical issues that I would like to
follow up, mostly centering on. the sections dealing with the
special character of scientific knowledge. Perhaps points that
seemed obscure or paradoxical at the time· of first publication
will now be more familiar and comfortable. I would really wel-
come the challenge to return to those. The whole book is an
exercise in the complementarity of philosophical and social criti-
 Introduction to the Transaction Edition xxiii
cism; perhaps from a renewed engagement with those very fun-
damental problems of scientific knowledge I will then achieve
an enriched perspective on its social problems.
J.R.R.
June 1995
 PREFACE

THE thoughts that are developed in this book have occupied me for
such a large part of my life that to describe their origins, or even to
give credit to all those who have influenced my progress, would re-
quire an autobiographical volume. My concern for the problems of
society and ethics has been with me since childhood; and my
curiosity about the nature ofscientific knowledge was aroused at high
school. On the former problem, I at least had Lincoln Steffens'
Autobiography as an illuminating introduction; but on science, I
found the accepted 'philosophy of science' to be only another
university subject, interesting in its own right but irrelevant to the
problems that were germinating in my own mind. For some years I
had hoped that Marxism would provide a coherent synthesis of all
these problems; but in the versions then extant it was very much less
a successful docttine than an extremely general guide to action. I
recall that I was aware ofa philosophical problem-situation to which
I wished eventually to devote myself some twenty years ago when
I first came to England from America; about seven years ago I had
an opportunity to make a first statement of the problems; and since
then my work on this material has become steadily more intense, up
to the present. Through all these years, I have been mulling over
problems in my mind; and they would gain evidence from experience
in a variety of ways, mostly unsystematic and informal. My reading
and teaching in the history of science was fundamental for the
growth of my ideas; indeed, had I not been able to become an
historian rather than a mathematician, it is very unlikely that this
book could have been written. Of equal importance, however, were
the insights and instances derived from a great multitude of dis-
cussions, with friends inside and outside the academic world. Many
of these discussions were on such ordinary and practical questions,
that neither I nor my friend was aware at the time that something
\vould be 'going into my book'; and when in retrospect I would
realize that a particular idea had been suggested or stimulated by a
discussion, it was often difficult to remember which one, with which
 xxvi Preface
friend, had been the occasion for the crystallizing of a vague notion.
As I have written up the text, I have tried to give credit to friends
for points which struck me at the time of discussion; I fear that this
biases the credits towards recent years, and also omits a goodly num-
ber of friends from whom I have gained a generalized wisdom.
It is possible for me to identify those institutions and organiza-
tions to which I am indebted, in various ways, for the growth of my
work. When I attended Central High School, Philadelphia, there
still remained enough of the atmORphere, and the teaching staff, of
the Gymnasium that it had once been, for me to experience the
challenge and discipline of serious academic study. Through the
generosity of the Pepsi-Cola Company in awarding me a scholar-
ship, I was able to attend Swarthmore College, where windows
opened for me. There I discovered the delights of intellectual ex-
ploration; and also found a community where people are honourable
and trustworthy in the absence of any propaganda or sanctions,
simply because of an abnosphere created by the College's Q!Iaker
traditions. Coming to Trinity College, Cambridge on a Fulbright
Scholarship, I had the privilege of apprenticeship to a master-
craftsman of mathematics, Professor A. S. Besicovitch, F.R.S. At
the (then) Durham Colleges in the University of Durham. I found
myself the protege and friend of the Warden, Sir James Fitzjames
Duff; and through him learned more of the subtle differences be-
tween English and American society. There too I met Mr. Peter
Doris, who taught' me the importance of finding the fundamental
questions and sticking to them.
When I moved to Leeds University to work with Professor
Stephen Toulmin on a Leverhulme Fellowship, I could concentrate
my energies on the history ofscience and the philosophical problems
suggested by it. The occasion for the first formal presentation of the
ideas of this book was provided by the Workers' Educational
Association; during 1962-3 I held a short course on 'science and
society', whose participants included colleagues and students from
my department and from the Department of Social Studies, and
secondary school pupils with whom I had worked in the Campaign
for Nuclear Disarmament. Just a half-century earlier, W.E.A.
teaching in Leeds had been the occasion for the achievement of a
most influential book in the philosophy of science, Norman Camp-
bell's What ;s Science? If this book is successful, it will serve as
another example of useful ideas being developed outside the context
 Pre/fIG' xxvii
of teaching to a captive audience preparing for degree examinations.
For several years, a small but vigorous discussion group held under
the auspices of the Leeds University Union Student Christian
Movement examined drafts of the basic chapters of this book, and
subjected them to a sympathetically ruthless criticism. Another
small but vigorous group, a seminar in 'Social Problems of Con-
temporary Science' at the Harvard University Summer School in
1967, helped me formulate the problems discussed in the latter part
of the book. Equally important for this work has been my contact
with the peace movement in Britain, and more recently, with the
British and American societies for social responsibility in science.
Even with all these opportunities and stimuli, I would not have been
able to create this book, and give it time to mature, were I not work-
ing in the happy and encouraging environment of the University of
Leeds. My debt to the University, through successive Heads of my
department and Vice-Chancellors, together with the administrative,
clerical and teaching staff with whom I have worked, is enprmous.
The motto of the University is Bacon's 'et augebitur scientia'; I
hope that my own work' will both explain its deeper meaning, and
contribute to the fulfilment of its prophecy.
Among the individuals to whom my obligations are particularly
strong, I recall from earlier years the late ProfessorA. S. Besicovitch,
Professor C. Truesdell, Dr. D. T. Whiteside and Dr. (now Professor)
I. Lakatos, for insights into the nature of the work and of the know-
ledge achieved in the exact sciences. When Dr. D. S. L. Cardwell
and I were colleagues at Leeds, we had many conversations whose
fruits are inadequately acknowledged in the occasional citations in
the text. On the basis of those discussions, I was able to appreciate
the insights contained in Michael Polanyi's Personal Kno1lJledge,
particularly in connection with the nature and role of craft skills in
scientific work. More recently, conversations with Dr. P. M. Rattansi
provided me with the first insights on the 'craftsman's romantic
philosophy of nature' and on the category of folk-science; and from
]. E. McGuire and C. Webster I learned much concerning the
philosophical and social aspects of science during its formative
period in the seventeenth century. To R. G. A. Dolby I am indebted
for his truly. incisive criticisms of crucial sections of the text. Dis-
cussions with Dr. N. D. Ellis and Dr. A. Coddington have been
essential for the development of my ideas on· 'immature sciences';
and N. Parry exposed the obscurities at the foundations of earlier
xxviii Pre/fIG'
drafts of my argument on ethics in science. Dr. Alison Ravetz has
contributed to the work oyer the years, through sharing her own ex-
perience as an archaeologist and historian, and also providing an
asttingent and penettating stylistic criticism at several crucial points
in the evolution of the text. It was through reading the work of
Professor Barry Commoner and his group at St. Louis, that I came
to the conception of 'critical science', which gives this whole work
what unity it has. I must also thank the two typists who produced a
fm text with such speed and enthusiasm, Mrs. Lily Minkin and
Miss Margaret Knapp; and also thank the proprietors of Beck Hall,
Malham, Yorkshire, whose friendly hospitality enabled me to
achieve an intense concentration on the most difficult problems of
shaping this book, in ideal conditions. I am grateful finally to the
Oarendon ·Press for their enthusiasm for this difficult book; and
to them and the printers for producing it so quickly and well. I
hardly need add that the many imperfections of this book, some
ofthem known to me but doubdess more unknown as yet, are entirely
my own responsibility.
July 197 1 J.R.R.
 INTRODUCTION

THE problem-situation that gave rise to the present work is the recent
rapid change in the character of the social activity of disciplined
inquiry into the natural world, and the consequent changes in its
understanding ofitself. The established traditions of research in 'the
philosophy of science' and in 'the sociology of science' have been
recognized as losing contact with the actual practice and real pro-
blems of science in the present period. Q!tite independendy of these
academic studies, there has developed a new common-sense under-
standing of science, derived from the daily experience of working
scientists and from the problems of decision and government within
science. I There has been a rapidly increasing flow of studies of one
or another aspect ofscience, conditioned by this new common-sense,
but none as yet that attempts a new synthesis.
It appeared to me that the problems of the character of scientific
knowledge, ofthe sociology and ethics ofscience, and ofthe applica-
tions ofscience to technology and to human welfare, are so intimately
connected that a proper study of anyone of them requires an in-
formed awareness of the others. A mixture of inherited assumptions
from the academic studies, with elements of the new common-sense;
varying in its contents from one person to the next, is not sufficient
for this. Hence I have tried to create a coherent framework of
I The most extensive codification of this new common ieDse of science is in W. O.
Hagstrom, Tile Stimtifie C"""""';'y (Basic Books, New York and London, 1965). Many
of the points made in this book can be found there, with • wealth of illustrative material
from interviews and historical examples. But his problem was very dif£el'ent from mine,
and the organization of the argument, and the emphasis given to the various points, is
quite difFerent. I felt that it would be otiose to provide a concordance of material between
the two texts, and so I am mentioning it at this early stage, with. reamunendation for its
study as a parallel source.
Rather closer to my own concerns are C. Wright Mills, Tile SotiolotittllI",.,;tllllitnl
(Oxford Univenity Press, 1959; Grove Press, New York, 1961) and D. S. Greenberg,
Tile Politics ofAmmcil" Sa,,", (penguin, London, 1969; first published IS Tile Poi,"s
of P", S""",, New American Library, New York, 196'7). I have quoted from them
occasionally to provide evidence for some particular points, but each in its own field
provides materials and insights complementary to those developed here.
2 Introduction
concepts in whose terms each of the problems could be discussed
in relation to the others.
I am well aware that my extended arguments can be difficult for
those who have not already begun to reflect on science in terms ofthis
new common-sense. In order to ease the way in, I have provided two
introductory chapters. The first shows that there is a problem in the
understanding of science. The second picks up several themes that
are current in popular discussions ofscience, and provides them with
a unified theme: the industrialization ofscience and its consequences.
In the main body of the text, the order in which topics are dis-
cussed is gov-emed mainly by the requirements of clearly defined
terms and established conclusions that are necessary for a proper
analysis at each point. I have therefore proceeded from the aspects
of science that are more abstracted from their social context, to-
wards those ofthe fullest complexity. Thus, it is possible to analyse
problem-solving in research with only an occasional reference to its
wider aspects, while the problems of quality-control and ethics in
science are difficult to appreciate without a systematic knowledge of
the craft ofresearch and the nature ofits products. Lest this ordering
from the simple to the complex give the impression that the earlier
topics are in any way more fundamental, I have referred forward,
informally, to later sections of the book whenever appropriate.
The sequence of major problems is then epistemology, sociology
and ethics, and applications in the technical and human context. On
the first, I am not so much concerned to refute the earlier assumption
that science somehow produces-truth by the application ofa standard
method, as to see how recognizeable, objective scientific knowledge
can (but need not) eventually emerge from the very personal and
fallible activity of research. On the second set of problems, I analyse
the sorts of social mechanisms and attitudes that are necessary in
order that worthwhile research may be done; and I also study the
conditions under which they are effective, and those under which
they are not. Having established the concepts required for a dis-
cussion of both the 'individual' and the 'social' aspects of scientific
work, I can enrich them further and discuss the applications of
'science' and 'the scientific method' both to disciplines concerned
with man rather than with nature, and to problems relating to tech-
nology and to human welfare. With all this accomplished, I can
return to the general problems ofthe present and future condition of
science, and offer a critical discussion of these. Because I build up
 Introduction 3
my technical vocabulary as I proceed, using ordinary words in special
restricted senses, it is not easy to dip into the book halfway through
and follow the argument. This difficulty is unavoidable, but the
analytical index should help the reader locate the main discussion of
the technical terms that I use in a special sense.
There is one basic term that I have deliberately used in several
senses, and that is 'science' itself: A glossary ofthe various meanings
of the English word now in use would be lengthy and tedious; and if
one extended this to the German Wust1lSchll{t and the French
science positive, both of which made important contributions to our
inherited notions of 'science', then any worthwhile study would be a
major undertaking in itself. Also, the different parts of my analysis
apply to different (though overlapping) senses of 'science', and the
context has been sufficient for establishing, in a general way, which
sort is being discussed. Thus in Part I, the concern is with the natural
sciences, pure and applied; while in Parts II and III, the analysis
applies to any sort of disciplined inquiry. In Part IV, the 'social
sciences', pure and applied, are to be included as a major object of
attention; and in the Conclusion, my concern reverts primarily to
the natural sciences. I have tried to indicate these particular focuses
of reference by my use of words standing for 'science' in the text,
and by the examples discussed in the foomotes. Although this
method still leaves some obscurity, it seemed preferable to a pedantic
differentiation of terms.
Another ambiguity in usage should be pointed out, as it may
eventually prove to be a weak point in my sttueture of concepts. I
have not considered the problem of the different levels of aggrega-
tion of problems, methods and results, described loosely by such
terms as 'field', 'discipline', and 'science'. There was no place
in my analysis where I could perceive any differeRce being made by
the level ofaggregation assumed. Hence I have rather promiscuously
used these terms as synonyms; and the occurrence ofone rather than
another should not be taken as having some particular significance.
In almost all cases, I have kept the discussion of particular ex-
amples in foomotes separate from the text. Although this may be
inconvenient for the. reader, and also deprives the examples of some
of their force as evidence, there are several reasons for my adopting
this style. The consistent alternative, of providing evidence for every
thesis by a significant example from history or current practice,
thoroughly established as valid in itselfand woven into the argument,
4 Introduction
would have required a German style of scholarship, involving a
preparation time and .a book both several times longer than the
present work. Also since historians of science have themselves only
recendy come to share the new common-sense of science that I am
trying to formalize, reliable and app.ropriate case-studies are still
rare; and to provide comprehensive evidence for each of my theses,
I would have needed first to re-create much ofthe history ofscience.
There are other reasons for the separation of examples from argu-
ment, deriving from the character of the work itself. No set of
particular examples can provide conclusive evidence for general
theses of the sort argued here. Many of my arguments will appear as
restatements of the obvious to those involved in the practice of
science, and for them the examples will function as light relief. On
the other hand, those who find my conclusions repugnant or in-
comprehensible would be resistant even to the most persuasively
argued case-studies; they could well dismiss each of them as un-
representative instances. Hence my arguments will carry conviction
if and only if they strike the reader as offering real solutions to real
problems; and for this, examples that illustrate the argument with-
out entangling and lengthening it seem most appropriate.
Although the main intended function of this work is a practical
one, to assist in the understanding and solving of the deep problems
confronting science now and in the near future, I have consistendy
attempted to maintain a standard of rigour and consistency appro-
priate to philosophical discussion. Also, I believe that some of the
concepts and distinctions I have introduced in the discussion may
be of use in other general discussions of science. The basic concepts
fall into two classes, those concerned with knowledge and those con-
cerned with action. For the former, I define 'problem', and give a
sketch of its anatomy, distinguishing the different stages of develop-
ment of its materials, as 'data', 'information', 'evidence', and 'con-
elusion'. In analysing 'methods', I distinguish among the criteria on
which the controlling judgements of scientific inquiry are based:
particularly those of 'adequacy' and of 'value'. I make strong use of
the concept of 'pitfall', which more than any other explains how
apparently rigorous research can lead to disasters. Studying the
subsequent history of a solved problem, I distinguish between the
'result' itself, the 'fact' which may evolve from it (and I define this
term closely), and 'scientific knowledge'. The condition of 'im-
maturity' of a field can be defined in terms of the absence of 'facts',
 Introduction S
resulting from the lack of those appropriate criteria of adequacy
which steer the investigation away from pitfalls.
Concerning action, I make strong use of distinctions within the
'final cause' of a task, such as the 'goal' which defines what is to be
attempted, the 'function' to be performed by the successfully
accomplished task, and the 'purpose' in human satisfaction to which
the performed function is a means. The operator on the task has his
own pwpose (or purposes) to be achieved, not necessarily identical
with the fulfilment of the goal; and for brevity I use 'motive' for his
reasons for being on that task at all. The application ofthis battery of
final causes to the social activity of science enables an analysis of the
problems of quality-control and ethics. Also, with it one can dis-
tinguish between the 'scientific problem' of the traditional sort, the
'technical problem' in which the goal is defined by the desired per-
formance of a pre-assigned function, and the 'practical problem'
defined by the achievement of given purposes. One can also use these
distinctions for an analysis of various pathologies of social activities
in general; in their terms I am able to define 'corruption',
Finally, all these concepts can be used to enrich the earlier dis-
cussion of important styles of scientific activity. The 'industrialized
science' of the present can be distinguished from the 'academic
science' that dominates the folk-memory of leading scientists of the
older generation, in terms of the capital-intensity of the tools of
scientific work and the consequent new social relations within the
world of science. We now see emerging a 'critical science', in which
science, technology, politics, and ultimately the philosophy ofnature
are involved, and which may be the most significant development in
the science of our age.
 Part I
THE VARIETIES OF
SCIENTIFIC EXPERIENCE
 INTRODUCTION

THE activity of modem natural science has transformed our know-

ledge and control ofthe world about us; but in the process it has also
transformed itself; and it has created problems which natural
science alone cannot solve. Modem society depends increasingly on
industrial production based on the application of scientific results;
but the production of these results has itself become a large and
expensive industty; and the problems of managing that industty,
and of controlling the effects ofits products, are urgent and difficult.
All this has happened so quickly within the past generation, that the
new situation, and its implications, are only imperfectly understood.
It opens up new possibilities for science and for human life, but it
also p(esents new problems and dangers. For science itself, the
analogies between the industrial production of material goods and
that of scientific results have their uses, and also their hazards. As a
product of a socially organized activity, scientific knowledge is very
different from soap; and those who plan for science will neglect that
difference at their peril. Also, the understanding and control of the
effects of our science-based technology present problems for which
neither the academic science of the past, nor the industrialized
science ofthe present, possesses techniques or attitudes appropriate
to their solution. The illusion that there is a natural science standing
pure and separate from all involvement with society is disappearing
rapidly; but it tends to be replaced by the vulgar reduction of
science to a branch of commercial or military industry. Unless
science itself is to be debased and corrupted, and its results used in a
headlong rush to social and ecological catastrophe, there must be a
renewed understanding of the very special sort of work, so delicate
and so powerful, of scientific inquiry.
If we are to achieve the benefits of industrialized science, and
avert its dangers, then both the common sense understanding of
science and the disciplined philosophy of science will need to be
modified and enriched. As they exist now, both have come down
from periods when the conditions of work in science, and the
practical and ideological problems encountered by its proponents,
10 TIJe VII,ietieS ofSeimtijie Ezperimte
were quite different from those of the present day. Science is no
longer a marginal pursuit oflittle practical use carried on by a hand-
ful of enthusiasts; and it no longer needs to justify itself by a direct
answer to the challenge of other fields of knowledge claiming
exclusive access to ttuth. As the world of science has grown in size
and in power, its deepest problems have changed from the epistemo-
logical to the social. Although the character of the knowledge
embodied in a particular scientific result is largely independent ofthe
social context ofits first achievement, the increase and improvement
of scientific knowledge is a very specialized and delicate social
process, whose continued health and vitality under new conditions
is by no means to be taken for granted. Moreover, science has grown
to its present size and importance through its application to the
solution of other sorts of problems, and these extensions react back
on science and become part of it. For an understanding of this
extended and enriched 'science', we must consider those sorts of
disciplined inquiry whose goals include power as well as knowledge.
To achieve an understanding of the new social character and
problems of science, it is useful first to review the deficiencies of the
various prevalent common sense and traditional images of science;
and then to show by example that the industrialized science of the
present has social problems which arise from the technical and
social conditions ofits work. This discussion opens the way to a new
philosophy of science which, instead ofasking 'What sort of ttuth is
embodied in perfected scientific knowledge?', proceeds by asking
'By what activities and judgements, individual and social, can
genuine scientific knowledge come to be?' The second part of this
book is concerned with this question, and the concepts developed
there are then applied to some important problems in the social
activity of science. In the third part I examine the practices and
attitudes which are necessary for the maintenance of the health of
scientific endeavour, and to prevent its decline into stagnation or
corruption. In the fourth part, I consider the special features of the
application of science to technical and practical problems; and in
this connection I analyse the ~ly difficult problems of
immature sciences. Finally, I discuss the moral and political prob-
lems of industrialized science, and show that it is already giving rise
to its opposite, 'critical science', which may offer a means to the
resolution of some of the problems which disturb and perplex all
those who recognize the condition of science in our modem world.
 I

'WHAT IS SCIENCE?'

THE SUCCESS and vigour of scientific research, and the effectiveness

of its technological application, are generally accepted as indicators
of the quality of a nation's life. Science is so important, and expen-
sive, that the major policy decisions concerning its development are
increasingly being taken by the State, rather than being left to the
judgement of the scientists and their private patrons. Accordingly,
the progress of science becomes a matter of politics; everyone in the
community is involved in the consequences of decisions on 'science
policy', and every citizen is responsible, however indirectly, for the
formation of those decisions. This social involvement of science will
necessarily increase over the decades to come. An increasing number
of practical problems will need to be solved through planned pro-
grammes of scientific and technological research, and the rising cost
of the many-sided work of science will call for the most direct
involvement of the State in its planning.
In such a situation, it is most important that there should be a
common knowledge of what it is that is being discussed and
planned, and what are the special features that distinguish it from
other objects of public concern. Such a common understanding
should eventually develop, but we are as yet a long way from it. At
present, we find a variety of partial views on the nature of science.
Each of these derives from some special experience of a very com-
plex activity, and from traditional views of science that were de-
veloped in response to problems that have since vanished. Such a
variety is natural; but the present situation has the potentially dan-
gerous feature that there is no general awareness of this variety.
Each view of'science' seems obvious common sense to its proponent;
and in discussions of science, one can find the participants using the
same word to refer to radically different things.
Thus, the common sense of science actually includes several
12 TIJe Varieties ofScientijie Ezperimte
varieties of scientific experience. The understanding of science is
truly fragmented; but each of these fragments reflects an important
aspect ofthe total thing. It would be im~ible to fit these fragments
together into a coherent whole, as if they were pieces of a jigsaw
puzzle; but any attempt at unified understanding of science must
take account of the experience (of the past and of the present) from
which these views take their origin.

Vie11JS ofscience
Since natural science depends on the general public for its support,
in this period as never before, the public understanding of science is
crucial, in the long run, for the continued health of the community
of science. In many ways, the public appreciation of science leaves
nothing to be desired. We are all grateful for the comfort and
security of life that is achieved by modem technology, and prepared
to accept the claim that all these good things are the by-products of
scientific research. Moreover, in the common-sense understanding
of man's relations with his natural environment, and even with his
fellows, 'science' reigns supreme. Supernatural explanations of
natural events, even of great disasters, are no longer taken seriously.
The areas of ordinary life where inherited craft-wisdom is valued
more highly than the judgements of scientific experts, are shrinking
down to the vanishing point. This arises partly from the increasing
artificiality of our material environment, and from the rapid changes
in it as well as in our social environment. A sign of the triumph of
'science' is the reliance on textbooks for such personal crafts as
rearing children and even achieving a happy married life. This
tendency is more marked in the United States than anywhere else;
there the prevailing assumption is that every problem, personal and
social as well as natural and technical, should be amenable to
solution by the application of the appropriate science.
Without such a general appreciation of science, it would be
impossible for the scientific community to continue to grow, and to
lay claim to an ever-increasing share of the national wealth. But the
limitations of this appreciation must be understood if the support is
to continue on a genuine and healthy basis. First, what the public
appreciates is not the same as that for which the support is required.
In all the varied activities of science (including popularization,
teaching at all levels, technological application, and administration
and decision-making) it is research that lies at the nerve-centte.
 'What;s Science?' 13
Without research-the continuous, disciplined advance from the
known into the unknown-these other activities would either lose
their meaning) or become stale, sterile and eventually corrupted. Yet
scientific research is most difficult, if not impossible, for a layman to
comprehend. The work itself cannot be appreciated without some
prior familiarity with the very specialized activity of investigating
problems in the context of an abstract, technical discipline; while
it has some similarities to the running of a business, or even to the
driving of a bus, its special character can be communicated only to a
layman with a well-developed imagination. And the object of the
research work, the new scientific result, is almost always so techni-
nical, so dependent for its meaning and significance on the context
of the existing information in the special field, that the layman can
never get more than the most general idea of what it is all about. In
this respect, the 'layman' may equally well be a scientist in another
field, for the problem of internal communication within the world
of science is certainly severe. But at least one working scientist
can recognize the moves in an exposition by another, in spite of
not understanding the content; while the man in the street would
find a detailed account of a new discovery to be totally incompre-
hensible.
Hence what the general public appreciates in science is not what
the scientists are doing. Rather, it can be classed under two headings:
techniques and natural magic. The first is the collection of devices
that make life easier to live, or the destruction of life more efficient.
The second is the production of strange and wonderful effects with-
out recourse to supernatural agents. Although the term 'natural
magic' fell into disuse some time ago, it is quite natural for this
interpretation ofscience to persist. Whet! the layman (child or adult)
sees a demonstration of some astonishing effect, can he imagine the
human endeavour and intricate technical work which led to its
creation? No; it is the effect itself which captures the imagination,
and produces wonder and delight: this is natural magic. Of course,
the audience for the 'wonders of science' is told that these achieve-
ments are simply the application of the laws of Nature, and that the
creator of the effect is in no sense a magician. But since the audience
cannot understand the laws ofNature that are relevant to the produc-
tion of the effect, it is strange and wonderful (until it becomes a
commonplace part of the environment) and so differs little from
natural magic of the old sort. We can see the strength of this attitude
14 The VII,ietieS of S'ientific Ezperitn&e
from the great popularity of projects that are politely described as
science but which are clearly recognizable as nearly pure natural
magic; the manned exploration of sPace is the best example.
In earlier ages, when natural science was pursued by a small band
ofenthusiasts with the support ofa handful ofpattons, the ignorance
ofthe general public did not usually matter. But it is now the general
public which, indirectly or directly, pays the piper, and which,
through its elected representatives, ever more frequently wishes to
call the tune. No amount of popularizing effort will overcome the
fact that at the present stage of civilization scientific research is a
rather esoteric activity. Even in their relations with their pattons in
the past scientists have sometimes been tempted to play something of
a confidence game, inventing a middle ground between what was
important to their pattons and what was worthwhile to them. The
temptation is even stronger now, but it must be resisted. If relations
between the scientists and the general public are to remain healthy,
they must be based on mutual respect between people in different
worlds. The scientist must not dismiss the layman as an ignoramus,
and the layman must ttust the scientist as a man whose work, while
incomprehensible, is genuinely worthwhile for its own sake. Such
mutual respect cannot be achieved until both sides recognize the
difficulties in the way of genuine communication; and of course
there is a special responsibility on the scientist to be worthy of the
trust placed in him.
Between the general public and the professional scientists lies a
rather large group, roughly between five and ten per cent of the
population, who have some detailed acquaintance with scientific
knowledge, although no experience ofresearch. These are the people
who have made a disciplined study of some branch of science;
students, present or past. It is widely hoped that a higher proportion
of places ofresponsibility in public life will be taken by such people,
so that there will be an intelligent and competent mediation between
the scientific community and the general public. In the student's
view ofscience there is assuredly little ofmagic; but unfortunately it
is faIse to suppose that the experience ofleaming a selection from the
body of existing scientific knowledge confers any insight into the
process of its creation.
Up until very recently the overwhelming impression of science
imparted to a student by all but a few eccenttic and gifted teachers
was that of a mass of accomplished, solid facts. The student's task
 'What;s Science?' IS
was to assimilate a pre-sorted sample of these and to be capable of
reproducing them under examination conditions. The facts could be
of many types: special bits of information, techniques of manual or
mental manipulation, and general laws and theories. Whatever
their sort, they were recognizably hard, and announced to be clear.
For a student in such an environment, the frontiers ofscience would
be located on the remote heights of a pyramid of facts, and the work
of research would be imagined as the addition of some more hard
facts on top.
Much of the teaching of science in the later years of secondary
schools and in universities is still conducted within this framework.
But in recent years there has been a massive onslaught from the
proponents of the many reformed syllabuses for the teaching of
science. The difficulties encountered by these reforms may serve to
remind us that while this older view of science is dangerously one-
sided it is by no means without foundation. The great achievement
of the natural science of the past few centuries, unique in the history
of civilization, is the wealth of facts (of all the sorts I mentioned
above) which it has produced. And these facts are, to an astonishing
degree, hard and reliable, and are usually sufficiently clear to serve
their purpose. Moreover, scientific research, even of the most
inspired and revolutionary sort, is not accomplished by a great man
opening his eyes to the world about him, but necessarily grows out
of the matrix of a body of highly technical special results. Except in
the very rare cases when a whole field is created, it is on the basis of
old facts that new ones are made. Finally, to be able to do this work,
the scientist must be an accomplished craftsman; he must have
undergone a lengthy apprenticeship, learning how to do things with-
out being able to appreciate why they work.
Thus, the problem of imparting a correct view of science to
students cannot be solved by throwing out the dry old facts and
bringing in the excitement of open-ended research. Some sorts of
students may enjoy this much more, and those who go out into the
world will have a better appreciation of research than students of
earlier generations trained under the former discipline. But one can-
not be a craftsman unless one can manipulate one's tools; and one
cannot appreciate craftsmanship in others, as in judging- between
solid and spurious research, unless one has been trained up to it one-
self. Thus the students of the present day may have either one of two
diametrically opposed views of science, depending on whether they
16 The VII,ietieS ofScientific Experience
have been the subjects of pedagogical experimentation. Coming out
of the old school, they will tend to see science as the mountain of
facts; and coming out of the new school, they will tend to see it as
questions with ends opening in all directions. Some of the new
syllabuses manage to strike a balance; but in the difficult and
delicate task of training inexperienced and immature minds to
appreciate the complex interplay offacts and problems it is too much
to expect uniform success within a short time.
It might seem that the group of people whose view of science is
the most important and valid are the working research scientists
themselves. But although research is at the centte of scientific
activity, it remains as one very specialized part ofa large and complex
whole. The experience ofthe research worker will, in its own way, be
as specialized as that of the student. His task is to achieve new
results in a special field; even if he teaches part-time (as at a univer-
sity) there will usually be little or no connection between the new
results he is creating and the established knowledge which he is
passing on. Also, unless he is already in a position of seniority and
responsibility, he will have little involvement in the work which
requires him to see his own efforts in the context of the field as a
whole, in making the judgements and decisions which determine the
directions of future research.
Over the generations there has developed a view of science which
is well suited to produce the intensely committed, completely
specialized activity which modem science has required from the
majority of its practitioners. For them, 'research' is not merely the
centre of scientific activity; it is science. Even if this is not stated
explicitly, it can be inferred from the attitudes towards those
involved in other aspects of scientific work. Those who have tried
research, but abandoned it for easier things, are objects of some
contempt; those who have never even ttied, but are merely teachers,
are to be pitied.
This exclusive, sometimes fanatical, concentration on the produc-
tion of isolated scientific results can be traced back to the German
scientific research schools of the last century. It is conducive to the
production of solid work, but it is also anarchic and selfish. Its
prevalence has discouraged the growth to intellectual maturity of
many scientists who would otherwise have been capable of it. It
creates personal difficulties for scientists who, in their middle years,
would be more happily and productively employed on other sorts of
 'What ;s Science l' 17
work related to science; but they are reluctant to have it said ofthem
that 'he has given up research'. The conditioning produced by years
ofconcentrated and narrow research has the effect ofmaking it more
difficult for the leaders of science to see the work of themselves and
their colleagues in its broader context, both within science and in
relation to the outside world. And finally, any social activity which
depends for its recruits on a supply of single-minded enthusiasts is
dangerously vulnerable to changes in intellectual fashion.
Science as the Pursuit ofTruth
The question 'What is Science?' supplies the title or the subject-
matter of many books on the 'philosophy of science'. In them the
question usually takes the form, 'What sort of truth is embodied in
completed scientific knowledge ?' Ideas developed in the course ofan
attempt to answer such a question will not be well suited for descri-
bing science as a human activity, always changing and never perfect.
Treatises on 'Scientific Method' written within such a framework of
ideas seem to have little relation to the real work of discovering new
knowledge and are frequently scorned by practising scientists who
have become amateur philosophers of science.
It would be a serious mistake to dismiss all this effort as irrelevant
to the understanding of science. The question itself is one of deep
and perennial philosophical problems, considered in the special con-
text of modem science. The essays in 'scientific method' are an
attempt to reveal the anatomy of scientific knowledge, analogous to
the way that Aristotle's logic of syllogisms functions for correct
verbal arguments. Moreover, many of the greatest 'philosophers of
science' have been scientists of distinction who were able to reflect in
a disciplined way on what they were doing. Even if this approach to
the question seems to provide few ofthe answers we are now seeking,
it will help our understanding if we appreciate why it once seemed
the best way, and also why it no longer does so.
The 'scientific revolution' of the seventeenth century, although
better described as a reformation in natural philosophy, had some of
the essential characteristics ofwhat we now recognize as revolutions.
In particular, it was inaugurated by a small group of prophets, fully
conscious of their role in attempting to destroy an existing order and
to replace it by something better. Their purpose was the achieve-
ment of truth, and they were committed to the view that this could
be achieved only through a certain approach to the study of the
18 The Va,ieties ofScientific Experience
natural world. A classic expression of this ideology can be found in
Galileo:
If this point of which we dispute w~e some point of law, or other part
of the studies called the humanities, wherein there is neither truth nor
falsehood, we might give sufficient credit to the acuteness of wit, readiness
of answers, and the greater accomplishment of writers, and hope that he
who is most proficient in these will make his reason more probable and
plausible. But the conclusions of natural science are true and necessary,
and the judgement of man has nothing to do with them. I
Although the men who consolidated and extended the scientific
revolution were generally neither so certain about achieving truth,
nor so exclusive in their claims for natural science, this Galilean
commibnent remained in the ttadition, available for renewed
emphasis in times of institutional and ideological struggle on behalf
ofnatural science. This occurred during the nineteenth century; and
a hint of its consequences for the self-consciousness of science, and
the philosophy of science, is given in the preface to the first edition
of Karl Pearson's classic Grammar of Science of 1892. There he
mentioned the fact that science previously had to carry on a 'difficult
warfare with metaphysics and dogma'; and the exigencies of this
struggle explained, for him, the deliberate concealment of the
'obscurity which envelops the principia of science'. 2 Pearson saw
that this 'warfare' had distorted the self-consciousness ofscience and
had harmed its teaching; and considering the battles as won, he
thought it appropriate to exhibit the structure ofscientific knowledge,
in its weaknesses as well as its sttengths.
The 'warfare' to which Pearson referred was documented by other
writers of his time; 3 and those who had fought for science were not
necessarily anti-philosophical or irreligious. But in Germany there
had been a bitter institutional struggle for the establishment of
I G. Galileo, DWope 011 Tile Gr,., WorM Systems, in The Salisbury Tnnslation, ed.
G. de SantillaDa (University ofQUcago Press, 1953), p. 63. The word 'judgement' in the
Jut IeDtence is • traDsJation of'arbitrio' and, like all key words, it is loaded with meanings.
The other modern English translation, that of S. Dnke, DWope COII&emi", The TJI)()
ClliefWtJrlll SystlllU (University of California Press, 1953), uses 'will' (p. 53). The latin
root bas the ambiguities from which the English derives both 'arbitrate' and 'arbitrary';
'judgement' seems best suited to this context.
:a Karl Pcanon, Tu Gr.""., of StinIe, (1St edn., 18g2; Everyman, London, 1937),
P·3·
J A. D. White,.A HisttJr:! oflu W.f.e ofStinlee ",;IA TlleoloO ill ClJristnulom, 2 vols.
(18g6; reprinIcd Do~ PublicatioDs Inc., 19(0), is the classic in this literature. White
was President of Comell University and American Ambassador to Russia.
 'What;s Science?' 19
natural science as a discipline worthy of recognition as different
from, and independent of, the academic philosophy of the earlier
nineteenth century. And in England there had been a frequently
hysterical public debate over the theory of the evolution of the
human species, with ministers ofreligion denouncing the Darwinists
as the new Antichrist. What was at stake in these struggles was the
autonomy ofthe goals and methods ofinquiry in natural science; the
freedom of scientists to draw conclusions from the evidence they
considered relevant, and to ignore that which they considered
irrelevant (such as traditional Christian teaching, or accepted
metaphysical doctrine). In the last resort the struggle was over the
social functions and social responsibilities of natural science. Some
saw it as subordinate in its functions to some other sphere of
learning or work, in particular religion; and they considered that
scientists had a responsibility to work in such a way that both their
methods and their conclusions would reinforce the authority of
accepted religion rather than endanger it. The best defence against
this was, as usual, a good offensive; and the ideology of truth in
science, a truth independent from and perhaps superior to that of
philosophy or religion, performed a useful function in the struggle
for the freedom of science.
The struggle flared up briefly in the 1930s, when a group of
eminent British scientists (some native-born and others Continental
immigrants) saw a new threat to the freedom of science coming not
only from the racist persecutions of the Nazis but, more insidiously,
from the pressures on science in the cause ofsocial responsibility and
Marxist philosophy, enforced in the Soviet Union and urged by
British supporters of Soviet Communism. 4 The later episode of the
capture of Soviet genetics by T. D. Lysenko, who desttoyed a
brilliant school in the name of Marxism, seemed to confirm all their
fears. In reaction to this disaster, a distinguished American historian
of science used the Galileo text in a condemnation of the Marxist
approach to science and its history:
To suppose otherwise is to give the game away. It is to suppose that the
papal court was justified when it decided to condemn Galileo on the
ground that the Copernican system, however convenient as a mathematical
device, tended to unsettle men's minds and weaken authority and order.
It is to agree that the Soviet state was justified when it imposed
4 For a brief history of the 'Society for Freedom in Science' see Sir Solly Zuckerman
Scientists and W", (Hamish Hamilton, London, 1966), pp. 141 if:
20 The Varieties ofScientific Experience
Lysenkoism on the science of genetics as a theory comfortable to
Marxist dogma.... 5
Thus the Galilean commibnent to truth in science, seen as a neces-
sary means to the defence of its freedom, has continued down to our
own age, and will be expressed whenever the freedom of science
seems endangered.
Such a directly political motivation for the conception of science
as the pursuit of truth is only a part of a long and complex history.
Partly because of its indubitable successes natural science has
seemed for many decades to be the paradigm of genuine knowledge.
In a variety ofways, the pursuit ofscience, or a popularized image of
scientific work or scientific knowledge, has functioned as a religion
in substitution for, or in opposition to, the accepted religion of a
society. At the more practical level, the belief in the possibility of
attaining some truth that will live on after one's personal death, has
furnished a motive for selfless and dedicated work by many scientists;
and some schoolmasters, forced by circumstances to transmit what
is mainly a manipulative craft, derive self-respect and peace of mind
from the conviction that they are imparting clear and distinct truths
to their young charges. Since the pursuit of truth can appear to be
the most noble and harmless of human activities, this conception of
science helps to justify their work to scientists themselves; and it
also helps them argue for public support, especially when public
relations work is done on behalf of expensive experimental fields of
research.
The Technocratic Conception ofScience
The obsolescence of the conception of science as the pursuit of
truth results from several changes in the social activity of science.
First, the heavy warfare with 'theology and metaphysics' is over.
Although a few sharp skirmishes still occur, the attacks on the free-
dom of science from this quarter are no longer significant. This is
not so much because of the undoubted victory of science over its
ancient contenders as for the deeper reason that the conclusions of
natural science are no longer ideologically sensitive. What people,
either the masses or the educated, believe about the inanimate
universe or the biological aspects of humanity is not relevant to the
5 C. C. Gillispie, letter to Ameri"." Scientist, 45 (September 1957), 266A-'74A,
answering aiticisms of his review of J. Needham Scimc, lind Civiliulion in Chi".,
voL ii.
 'What is Science l' 21

stability of society, as it was once thought to be. The focus of

sensitivity is now in the social sciences; and the techniques of
control by those in authority vary in subtlety in accordance with
local requirements and traditions. Hence the leaders of the com-
munity of natural science no longer need to hold and proclaim this
sort of ideological commitment as a rallying slogan for their followers
and potential recruits. Also, the experience of modem scientists in
their work, seeing the rapid rate of obsolescence of scientific results,
makes the vision of the pursuit of truth not so much wrong as
irrelevant. But, more important, the attention of the general
educated public has shifted away from the problem of the nature of
pure science and its relation to philosophy and religion. It is now
concenttated· on the visible triumph of technology based on applied
science. Applied science has now become the basic means of produc-
tion in a modem economy. 6 The prosperity, and economic indepen-
dence, of a firm or of a nation does not rest so much in its existing
factories as in the 'research and development' laboratories, where the
industry of the future is being created and the competition of the
future is being met. Thus, industry has been penetrated by science.
All this has not only happened but it has recently been seen to
happen by those who plan the future of our societies. The 'techno-
cratic' view of science is that ofa basic factor of production, needing
ever-increasing supplies of highly-trained 'scientific manpower'.
This view of science is a descendant, in a simplified and vulgarized
form, of a tradition extending from Francis Bacon down through
Karl Marx. Bacon gave the aphorism 'knowledge and power meet in
one'; and Marxist historians have attempted to show that the major
advances in science have come as a response (however indirect) to
the particular needs of production at the time. Even in the nine-
teenth century, 'science' was given much credit for the advances in
technology which so dramatically transformed life in the advanced
nations; and so it is the widespread adoption of this view, rather
than the insight itself, which is characteristic of the present age.
But the 'technocratic' view of science is dangerously one-sided. It
assumes 'science' to be an independent, self-contained factor in the
situation, needing only financial support and administrative planning

6 Science can thus be considered as one sector of the 'knowledge industry'. One study
of this (F. Machlup, The Production lind Distribution ofKno7IJledge in the United SllItes,
Princeton, 1962) estimates its product at 27 per cent of the G.N~P.; although his
definition of 'knowledge' would be better termed 'software'.
22 The VII,ietieS of Scientific Ezperitnce
to provide unlimited blessings for us all. Until very recently, there
was no systematic appreciation of the fact that as science grows and
penetrates industry it becomes industtialized. This is partly a matter
of scale; as the invesbnent in scientific research increases, both as a
whole and in the cost of individual projects, there arises a need for a
formal administrative machinery for decision-making and executive
action. Moreover, many practically useful scientific results can be
tteated as a sort of commodity, to be produced under contract,
tailored to the special needs of the purchaser. The bulk of the public
funds which are given to 'science' are in fact contracts for such
specialized products; and it is only natural for the older forms of
pattonage of science to be displaced by contractual agreements for
the production of results, whether directly useful or not. Hence, in
any large and vigorous field, the social atmosphere becomes increas-
ingly 'industtial' where a large organization, with labour force
directed to specialized tasks, produces the sorts of results for which
the directors have been able to obtain contracts from agencies which
invest in such production.
Without some such organization it would be impossible for the
scientific community of the present, and of the future, to operate.
But the assimi1ation ofthe production ofscientific results to the pro-
duction of material goods can be dangerous, and indeed destructive
of science itself. For producing worthwhile scientific knowledge is
quite different from producing an acceptable marketable commodity,
like soap. Scientific knowledge cannot be mass-produced by
machines tended by semi-skilled labour. Research is a craft activity,
of a very specialized and delicate sort. The minimum standards of
accuracy and reliability for worthwhile scientific results are extremely
high. But there are no automatic, external tests of its quality, neither
gauges to check its agreement with specifications, nor a market
mechanism for the public rejection of inferior products. Two sepa-
rate factors are necessary for the achievement of worthwhile
scientific results: a community of scholars with a shared knowledge
of the standards of quality appropriate for their work and a shared
commitment to enforce those standards by the informal sanctions
the community possesses; and individuals whose personal integrity
sets standards at least as high as those required by their community.
If either of these conditions is lacking-if there is a field which is
either too disorganized or too demoralized to enforce the appropriate
standards, or a group ofscientists nominally within the field who are
 'What is Science l' 23
content to publish substandard work in substandard journals-then
bad work will be produced. This is but one of the ways in which
'morale' is an important component of scientific activity; and any
view of science which fails to recognize the special conditions
necessary for the maintenance of morale in science is bound to make
disastrous blunders in the planning of science.
It may appear that this concern is somewhat beside the point. It is
applied science, or technology, that enters into the system of indus-
trial production, and every national plan for science ensures that
'pure science' or 'basic research' should not be neglected. Unfortu-
nately, it is impossible to make a neat line of division between the
two sectors, allowing one to serve Truth and the other Caesar. It is
more than the simple fact that no piece of scientific knowledge can
be guaranteed 'pure', or free of application. In addition, much of
'pure' or 'basic' research involves capital outlay on an industrial and
occasionally gigantic scale, and those who undertake and organize
such projects must necessarily have some of the qualities of an
industrial manager.
No sector of science is immune from the problems of industriali-
zation; and, conversely, modem sophisticated technology is suffi-
ciently like pure science to share its delicacy, and is also prone to
special diseases of its own. For technological innovation is also
removed from the immediate controls of the market. Every impor-
tant new device or process takes several years of development before
it can be produced, and hence a decision to invest in development is
necessarily partly a speculative assessment of prices, demand, and
competition in the future. In military technology, the criteria for
decision are even more complex and speculative. Mistakes can be
made here, as in science; but here they cannot be quietly buried.
There is then a danger that a firm (or more commonly a government)
will try to redress an error by making it bigger, and all considerations
of what is economically viable (not to say socially desirable) will be
cast aside. Supersonic airliners are a case in point. Hence scientific
technology, as much as industrialized science, can suffer from a new
and dangerous form of corruption.
The 'Humanist' Critique
Now that 'dogma and metaphysics' no longer do battle with
science, the traditional opposition views are collected together under
the label 'the Arts'. This is usually considered to be an outsider's
24 The Varieties ofScientific Experience
view, seeing science as a new Leviathan; a body ofknowledge which
is esoteric, inhuman, and increasingly dominant. Those who hate
and fear this new science are even more mixed than the Oxford
classicists and Chelsea poets who provided the initial evidence for
C. P. Snow's famous lecture.? Indeed, such critics are even to be
found within the camp of science itself, especially among students
who have been discreetly press-ganged into a scientific career by
their schoolmasters, or research workers in the position of helpless
proletarians on the production-line of some modem research
establishment.
The debate between the 'arts' critics and the defenders of science
is necessarily confused since it depends on a partial understanding on
both sides, and also involves a mixture of personal philosophy,
politics, and quite directly professional concerns (as among the
teachers of upper forms of English grammar schools). Also, the
debate oscillates between two levels: that of scientific knowledge
itselfand that ofits technological applications. This further confuses
the debate, for while the results of scientific inquiry are, in their
statement, remote from specifically human concerns and values, the
industrial applications of those results are ttansforming the material
and social conditions of human life. A fully consistent 'humanist'
critic must not merely point out the sterility (in his terms) of know-
ledge of the natural world but he must also deny the human value of
modem scientific technology. This latter task can be done, but only
by fabricating a legend of some bygone 'good old days' when men
were poor but happy, or by denying the human importance of the
majority of the human race. The fully consistent defenders of
science also have their difficulties. It is easy enough for a certain sort
of intellect to find only ttiviality in the works of literature and
philosophy, and to claim that natural science has the only path to
truth. 8 But the social responsibilities of scientists, in the face of
, c. P. Snow, Tile T"", CtdMes lind lile S&imljjie Ref}(JIIIIitm, Rede Lecture, 1959
(Cambridge University Press, 1959).
• Thus, I. I. Rabi, 'Science and Technology', in Tile I",PIIt' ofSeimee tm Tec/mology,
eels. Warner, Morse, and Eichner (Columbia University Press, 1965), pp. 9-36, can say
'The arts are certainly central to life. Yet they are not the kind of thing that will inspire
men to push on to new heights •.• Even the works of Shakespeare, which are essentially
an exploration of human character, are really wonderful, glorified gossip. I am not
degrading gossip, because we live by it, but it does not take us outside ourselves, outside
the human race' (p. 20). In fairness to Professor Rabi, it should be added that his
c:omments on social aspecII of science, in his essay and in the subsequent discussion, are
• valuable IOUI'CC of information and insights.
 'What is Science l' 2S
destructive and evil applications of their work, cannot be brushed
aside lightly. A tempting way out is to tteat the scientist's search for
truth about nature as an absolute good, and then to disclaim all
responsibility for the application of this knowledge at the hands of
the technologists. In an age when industtialized science and
science-based industry are so closely related this neat bifurcation
may strike an unsympathetic audience as shallow, if not indeed
hypocritical. 9
Between these extreme attitudes lies a rich mixture of sincere but
worried criticisms and defences. The critics of natural science
implicitly (or explicitly) contrast it to their idealized image of Arts
studies as explorations in the philosophy of man. The teaching of
science is claimed to be irrelevant to true education, and indeed
destructive of the values of education. It is authoritarian in style,
giving students no opportunity to develop their powers of criticism
and judgement. Dealing with an artificial and abstract subject-
matter, it gives the student no materials for developing himself as a
mature person in human society. Moreover, run the criticisms,
scientific knowledge cannot be part ofthe intellectual equipment ofa
cultivated person. Either one is a highly trained specialist, or one can
learn of science only at third-hand through populariutions. The
specialists themselves are accused of having a narrow vision of the
world since their training and professional practice removes them
from contact with genuinely human concerns. The increasing power
of natural science thus threatens the destruction of a humane
understanding among educated people as the humane studies are
increasingly deprived of prestige, of time in university teaching
programmes, and of resources for research. As a result, our thinking
about ourselves and the world around us becomes grossly material
and quantitative; the higher sensibilities and values are crushed
beneath the machine.
This comprehensive indictment would have more force if the
aggrieved humanists had previously achieved reliable methods for
teaching wisdom through their various approaches. But as Plato
9 L. Ie. Nash, in Tile NfI'"e of,lIe N.,,,.r Seim&es(Little Brown & Co., Boston, 1963),
distinguishes between the mom responsibility of the sdmrist, when he 'elects to take
put in the technologic exploitation of science for destructive purposes', and sdmrifie
fto7l1letlge, 'ethically as neutral as iron', capable of becoming swords or ploughshara.
He concludes, 'Ambivalence attaches to the works of science simply because their
technologic exploitations rest in the bands of men' (p. 113), but he does not discuss
whether a similar ambivalence could be justifiably attached to scientists.
26 The VII,ietieS ofScimtijiG Ezperimee
showed long ago, the task is not an easy one. Descartes' savage
indictment of his own humanistic education at one of the very best
schools of Europe should be read by all those who believe that the
virtues of 'humanism' are so obvious as to need no defending. lo
Moreover, most ofthe present criticisms ofnatural science can apply
to any developed discipline, be it classical philology or Shake-
spearean criticism. For illumination on the great, eternal questions
can be achieved and transmitted only by those rare individuals whose
intellect and vision are equal to the task. The vast majority of
scholars must do the humdrum work of investigating narrow,
technical problems, and teachers must try to impart some compe-
tence to immature minds. When people trained as research tech-
nicians are forced to be philosophers in their teaching, the results
can be disastrous. The deepest and noblest ideas of great minds,
when retailed by lesser men, can easily become banal commonplaces,
mockeries of the original inspiration. It is also worth remembering
that on occasion some 'humane' disciplines have suffered from
precisely that dehumanization of which natural science is now
accused; it was most marked in classics and history in Germany at
the end of the last century.11
In fact, the differences between 'Arts' and 'Science' disciplines
are more of degree than of kind. Rivalry and conflict between
different spheres oflearning is as old as learning itself. The perceived
lines of division, and the sharpness of conflict, are the result of
temporary and local circumstances. It can be very entertaining to
proceed from impressionist judgements on a very partial experience,
by undisciplined generaliution to erect abstract entities like 'Arts'
and 'Science', and then let this pair of straw men do battle with each
other. But without a recognition of the many-mdedness of the work
of the advancement of learning in any sphere, and of the systematic
difficulties common to all, little new understanding will be
achieved.
10 In the Diseows tle I. M,,1uHIe, 1st part, Descartes fint describes his enthusiasm for
the ftrious subjects in his hUlDlDistic curriculum, and then disposes of them one by one,
showing that they either confuse or mislead the mind, or aetua11y corrupt our morals, or
(u theology) are condemned to sterility by their own uscrtioDs. Only mathematics
survived his destructive analysis.
I I See F. Lilge, 1711 Ainue of u . .",,· IIIe F.lw, of IIIe Cmna U";""'sily (Mac-
Millan, New York, 1948), elL 4: 'Criticism and Satire of Academic Culture: N'1dZBChe'.
This unfortunate philolopher was not alone in his aitiqUCj the great Swiss historian
Jacob Burckhardt also stood aside from the industry of DUTOW &ctologic:al historical
scboIanhip (pp. 100-4)·
 'Whllt ;s Science l' 27
It is worth recalling that the lecture which gave rise to the pro-
longed discussion of the 'Two Cultures' was really directed towards
quite a different problem. An earlier title was 'The Rich and the
Poor', and the purpose of the lecture was to condemn the ignorance
in both 'cultures' of a phenomenon which will determine our
future: the power of the new scientific technology for transforming
the lives of the men, women and children now helping to emerge
from what Snow calls 'the great anonymous sludge of history'. But
Snow's English audience was not prepared to comprehend the
challenge laid down in that thesis, and so the tea-cups rattled
agreeably while the revolution went on outside.

Science as Dirty Work

Q!1ite recently there has developed another very critical image of
science: not so much that it is destructive of human values, but that
science, through its military and industrial applications, may well
destroy the human race. There has long been a popular tradition of
imagining 'the scientist' as a white-coated magician, concocting
weird and potent substances in his fuming test-tubes; the comic-
book literature provides evidence for this. But this image has been
far from dominant, as we can see from the very limited overt popular
reaction to the early nuclear weapons. However, as the political
campaigns against the use and testing of nuclear weapons gained an
audience, and u~e world found itselfon the brink ofnuclear annihila-
tion over Cuba, the sorcery of military science became a part of
everyday common sense. Even this new image did not react directly
on the activity of science, except perhaps in contributing to the
beginning of the decline in recruibnent. But shortly thereafter the
problems of the environment, or 'biosphere', began to come into
prominence. At the same time as the public was learning ofever new
and more subtle ways in which their technology is poisoning the
environment and themselves, there arose several large movements
against the society ofwhich this technology is an integral part. These
ranged from protests against the Vietnam war and its particularly
sophisticated horrors, through 'student' revolts against various
manifestations of bureaucracy, over to the 'hippie' or 'drop-out'
communities which have abandoned our civilization entirely.
The tarnishing of the image of science is very real, and is likely to
be one of the main reasons for the international 'swing from science'
28 The Varieties ofScientific Experience
among schoolchildren. 12 It is certain to persist, especially since
distinguished scientists now voice their perplexity and disquiet over
the work they have been engaged in. 13 Yet to conclude that all
science is 'dirty work', and solely responsible for our present
ecological problems, is both naive and fruitless. It is true that our
runaway technology is bringing us towards the brink of disaster; but
old-fashioned myopic engineering was quite sufficient to do the same
(on a smaller scale) in earlier times. The salination of irrigated lands
and the deforestation of the countryside have contributed to the
decay of civilizations through all of human history. Moreover, the
problems of the biosphere will not be solved by protests alone.
Even to determine what sort of damage is being caused, and by
which agents, is a sophisticated scientific problem; and to devise the
means of rectification and control calls for a variety of talents. If the
study of science is abandoned by all except those who are content to
be manpower-units, then within a generation there will be no one
competent and concerned to analyse and expose the ecological
blunders produced by our bureaucratic technology. In that case, the
hippies will have been proved correct, and we can anticipate at least
some centuries of barbarism before the next try at civilization.
Tow/lrds /I New Understanding
Although none of the partial views of science are adequate for an
understanding of the industrialized science of the present, three of
them have great significance for the future course of science. The
ideal ofthe pursuit oftruth has functioned for generations to provide
science with a defence against its external enemies and with a basis
for the morale and commitment necessary for the production of
worthwhile work. It will persist for some time to come, but its base
in experience is eroding very rapidly. The clue to its demise lies in
12 There is evidence that the image ofscience as 'dirty work' bas now become diffiJsed
among younger schoolchildren; and to the extent that this is so, the problems of
recruitment to science cumot be expected to be .solved for nearly a decade. In a survey of
children of ages 12 to I], using relatively unstructured discussions, and picking up
repeated key words u clues to Steleotypes, the most often repeated theme (18 per cent of
the total against 10 per cent for any other) was along these lines: 'They may invent good
thinp like new drup •.• well, other thinp I can't name but also things which are not
very good (18 per cent) like H-bombs and other weapons, giving diseases to animals j and
the thousands of scientists breeding germs.' See C. SeImes, 'Attitudes to Science and
Scientists: the Attitudes of 12-13 Year-Qld Pupils', ScluJol ScinI&e Rmnl1, 51 (1969),
7-14-
13 For an eumple of this new self-questioninl, see S. Silver, 'Science and Society',
ScinI&e JOfInUII (October, 1969), pp. 39-44-
 CWhat ;s Science l' 29
the concept of'scientific manpower'; and a warning about that term,
issued some years ago by Sir Eric Ashby, can now be seen as the
epitaph of academic science and with it the ideal of truth.
I do not deny that for certain very narrow purposes it is convenient to do
sums in which scientists and technologists are considered as so many units
of scientific or technological man-power. The danger is that people who
habitually do such arithmetic come to think ofscientists and technologists
as nothing but man-power units, and the places which produce them as
assembly-lines for man-power production. It is the unforgiveable sin to
introduce the concept of man-power into education. For centuries
universities have been concerned with individual men, not man-power.
No one, except as a joke, talks about classical man-power or philosophical
man-power. Universities must be firm, even with their friends, in rejecting

I.
any implications that they are concerned with man-power. They are not.
They are concerned with individual men.
This was written in 1958, and since then the calculations with
units of scientific manpower have increased steadily. Indeed, the
term 'scientist' itself is giving way to the acronym ~S.E. (Q!talified
Scientist or Engineer) as a grading for a particular class of skilled
manpower. If Sir Eric Ashby were a backward-looking romantic
then his worries could be dismissed; but he is a scientist of distinc-
tion, with a broad experience ofthe world ofscience and universities.
That his 'unforgiveable sin' has now become standard practice is
evidence of the complete victory of the technocratic conception of
science.
Science lives not by manpower units alone; and without some
ideal to replace that of the pursuit of truth it could soon degenerate.
Moreover, to the extent that science becomes organized around the
service of commercial and military industry, it will be subject to the
criticisms of being dirty work. Attempts by leaders of science to
conceal this connection from an audience of the lay public and
potential recruits, while using it to obtain support from industry and
the State, will accomplish nothing but a further decline in the
prestige and morale of science.
There is certainly no easy way out of this impasse; and to the
extent that industrialized science depends on large-scale investment
from industry and the State, it must be responsible to those institu-
tions. But it is possible that the reaction against runaway technology
will not take a purely negative form. Many leading scientists have
14 Eric Ashby, Te&1mo101] .tUl tile A&fIdemi&s (MacMillan, London, 1958), p. 94-
30 The Varieties ofScientific Ezperimce
already begun to criticize what is done in the name of science; and
there are now emerging focal points of 'critical science' where
research is organized around the identification and exposure of the
damage being done to the biosphere. As yet this movement is small;
and, even if it grows, it will necessarily exist on the margin of
industtialized science. Also, it must expect incomprehension and
hostility when it tries to establish an institutional base for its
teaching; for by its programme it implicidy rejects our inherited
approach to science, with its faith in Facts and in Progress, as a relic
of the Victorian age. But as it gains strength and coherence it will
develop a new philosophy of science, and a new philosophy of
nature and ofman's place in it. For this, it can draw on a suppressed
tradition within natural science itself, which saw beyond the
accumulation of facts, and beyond the domination of nature, to the
welfare ofhumanity living in harmony with itselfand its neighbours.
Whether such a philosophy could flourish within the context of our
industrial civilization, and whether the new science based on such
a philosophy could gain influence in time to avert the destruction of
civilization, are unanswerable questions. But it is certain that without
a new understanding of science our present problems, within
science and in its applications, will multiply to the point ofno return.
 2

SOCIAL PROBLEMS OF
INDUSTRIALIZED SCIENCE

IN recent years, the vision of'science' as the pursuit ofthe Good and
the True has become seriously clouded, and social and ethical
problems have accumulated from all directions. Each particular one
of these can be discussed in isolation, and solutions sought as if it
were the only or the main perturbing force in an otherwise tranquil
scene. But there are systematic connections between these various
problems, which can be organized under the tide 'industtialization
of science'. This means, in the first place, the dominance of capital-
intensive research, and its social consequences in the concenttation of
power in a small section of the community. It also involves the
interpenetration of science and industry, with the loss of boundaries
which enabled different styles of work, with their appropriate codes
of behaviour and ideals, to coexist. Further, it implies a large size,
both in particular units and in the aggregate, with the consequent
loss ofnetworks ofinformal, personal contacts binding a community.
Finally, it brings into science the instability and sense of rapid but
uncontrolled change, characteristic of the world of industry and
trade in our civilization.
The social and ethical problems consequent upon the industriali-
zation of science are the deepest problems in the understanding of
the science of our period. In response to these problems, both the
informal common-sense understanding of science, and the scholarly
philosophy of science, have been shifting their perspectives and
modifying their assumptions about what is 'real' in science. The
present work is intended as a contribution towards the under-
standing of these problems; and before going into a lengthy discus-
sion ofthe foundations, it will be useful for us to review the problems
themselves. Much of what I say will be harsh and unpleasant, and
to some will appear to be muckraking for its own sake. But these
32 The Varieties of Scientific Ezperimee
things can and should be said, not only because they reflect the
judgements of eminent scientists, but also because there are many
men who have devoted their lives to science, whose efforts to solve
these problems will be helped by their identification and explana-
tion.
Paradise Lost
A very important feature of the social problems of contemporary
science is their novelty.. Although the gross quantitative indices
of scientific work show a fairly steady growth-rate over many
decades,1 the self-awareness ofthe scientific community, as reflected
in the pronouncements of leading men in leading fields and in the
sort8 of studies made of science, has experienced a sudden change.
Whether the situation of science as a whole has really altered so
drastically in the very recent period is not of crucial importance in
this context. For the belief in the rapidity of change means that the
present leadership of science considers itself as lacking a stock of
inherited attitudes and social skills for coping with the problems
created by the present condition of science. As we shall see, the new
awareness of social and ethical problems of great complexity,
together with a sense of changes beyond the control of the scientific
community itself, and of inexperience in handling them, are them-
selves factors in a complex and unstable situation.
The sense of loss of a very precious quality of pre-war science is
conveyed clearly in the reflections of the distinguished Soviet
physicist Kapitsa:
The year that Rutherford died (1938) there disappeared forever the
happy days offree scientific work which gave us such delight in our youth.
Science has lost her freedom. Science has become a productive force. She
has become rich but she has become enslaved and part of her is veiled in
secrecy. I do not know whether Rutherford would continue nowadays to
joke and laugh as he used to do. 2
The loss of innocence of science is even deeper than Kapitsa
indicated. Science is not merely harnessed to industry; it can be
I The pioneering work in estab6ahing these statistical regularities and in appreciating
their sipificance was done by D. ]. de S. Price. See his ullie Snmee, Bi, StinI&e
(Columbia University Press, 1963). I was informed by the author that the basic quantita-
tive regularities were discovered by the techniques of 'little science': working from data
obtained by IUDpliDg IlaDdard ~ works. The computers were brought in later.
2 Qpotecl from v. K. McElheny, 'Kapitsa'. VISit to England', Snmee, 153 (1966),

725-7·
 Social Problems ofIndustrialiutl Science 33
applied to produce effects which for some centuries, since the
decline ofbelief in magic and miracles, had been thought impossible.
Norbert Wiener described this revival of ancient moral problems
under the heading of simony: the perversion of the magic of the
Host to other ends than the Greater Glory of God.
Dante indeed places it among the worst of sins, and consigns to the
bottom of his Hell some of the most notorious practitioners of simony of
his own times. However, simony was a besetting sin ofthe highly ecclesias-
tical world in which Dante lived, and is of course extinct in the more
rationalistic and rational world of the present day.
It is extinct I It is extinct. It is extinct? Perhaps the powers ofthe age of
the machine are not truly supernatural, but at least they seem beyond the
ordinary course of nature to the man in the street. Perhaps we no longer
interpret our duty as obliging us to devote these great powers to the
greater glory of God, but it still seems improper to us to devote them to
vain or selfish purposes. There is a sin, which consists of using the magic
of modem automatization to further personal profit or let loose the
apocalyptic terrors ofnuclear warfare. If this sin is to have a name, let it be
Simony or Sorcery.3
Still more serious is the changed perception of the character of
scientific work, and of the people who engage in it. For in the
traditions of propaganda on behalf of science, as well as in the
practice of the social activity of scientific inquiry, the work requires
codes of behaviour which are more sophisticated and more enlight-
ened than those adequate in many other spheres, and hence the
scientific community requires members ofsufficient moral calibre to
adhere to them. This ethical aspect of science has long been an
important part of its self-awareness. Since science cannot honestly
be defended purely on its contribution to industry, and its products
are generally too esoteric to contribute to general culture even at the
highest level, the work of scientific inquiry as a worthwhile thing in
itself must be an important component of any justification of the
total activity of science.
Traditionally, the image of science has been of a work which is
demanding, and also productive, of the highest standards of
morality. The work is arduous, involving the sacrifice of leisure and
convenience and lacking any of the rewards of wealth and power
which are supposed to spur ordinary mortals on in their endeavours.
In addition, scientists offer the products of their work openly and
3 N. Wiener, Coil tS Colnn, 111&. (M.I.T. Press, 1964), pp. 51-2.
34 The V.ties ofScientific Ezperim&e
without charge to all colleagues, ignoring the barriers of politics,
nationality, race, or class. Moreover, the inevitable debates over
scientific results are conducted within a set of rules which seem
inhumanly strict, and yet which are adhered to by consensus. The
scientist cannot hire an advocate to make his case sound better than
it is; he must not resort to personal abuse or denigration of oppo-
nents, or use his political influence (within the community or with
lay society) to determine an issue; and any resort to dishonest
reporting ofwork is sufficient to ruin a man's reputation. The willing
adoption of such rigoroUS standards of practice, and their operation
in the absence of any formal institutions for enforcement, seems to
require individuals of a genuinely superior moral character, who
willingly submit to such a discipline in the service of a noble end:
namely, the advance ofknowledge and power on behalfofhumanity.
There is no doubt that this elevated view of the ethical aspects of
science has had a strong foundation in reality, in the social practice
of science as well as in its public self-awareness, during the period
of'academic science' which began to be displaced during the Second
World War. The initial contact of scientists raised in pure academic
science, with men of affairs, often produced reactions like the one
expressed by the physicist Szilard:
When a scientist says something, his colleagues must ask themselves
only whether it is true. When a politician says something, his colleagues
must first of all ask, 'Why does he say it?'; later on they mayor may not
get around to asking whether it happens to be true. 4
The context of this quotation is a story called 'The Voice of the
Dolphins', in which an international organization of scientists has
found the simple secret of gaining the adherence of any politician to
any cause, irrespective ofhis predispositions: find his price, and buy
him. Szilard's estimated round figure at which any American
politician's soul could be purchased was a guaranteed lifetime
annual income ofS2oo,ooo (at 1960 prices).5
In recent years, the ttaditional portrait of the noble scientist has
receded from the centte of the image of science. There has not been
a denial of its accuracy for that part of scientific work which it
describes; but there has been a growing realization that there is more
to the activity of science than the achievement and assessment of the
4 See L Szilard, TM Voil, of 1M Doll_'" IIIIIl 01. Slories (Simon & SchUlter,
New York, 1961), pp. 35~.
5 Ibicl., p. 41, where the point is illustrated by example.
 Social Problems ofIndustrialiutl S'ience 3S
results of research. And the tangible rewards of a successful career
in high-prestige pure science, and in industrialized science, are
common knowledge. Szilard himself recognized that when scientists
engage in work other than straightforward research they are subject
to the same hazards as anyone else involved in practical affairs. In
another short story, he sketched a plan for achieving a desired
deceleration of scientific progress in the future through the creation
of an administrative machinery that would turn scientists into
bureaucrats. In this way, talented men would be diverted from
research, and mediocrity would be systematically fostered. His
model for this stifling bureaucracy was, very unfairly, the American
National Science Foundation. 6
Nostalgia for the bygone simplicities of particular fields of
science can exaggerate the depth and novelty of the social problems
ofcontemporary science, and thereby hinder their solution. Also, the
confusion among lay audiences between 'pure' and 'applied' science
and 'technology' obscures real differences which must be appreciated
if the different aspects of scientific work are to be understood and
maintained in a healthy relation. Yet this confusion is not altogether
unjustified; the creation of nuclear weapons was the achievement of
men who had been trained in the most pure and philosophical field
of science, atomic physics. And even 'pure science' has come to
require gigantic projects such as high-energy particle accelerators,
involving vast funds and also managerial and political skills for their
achievement.
In these new conditions, there has emerged a new literature about
scientists, in which the scientist is seen primarily as a man of affairs,
and the painstaking, disciplined work of research is only a part,
perhaps a minor part, of the story. There have been protests at this
practice of ignoring the science when talking about scientists;7 but
6 0p. at., pp. 89-102, 'The Mark Gable Foundation', especially pp. 100-1. The plan
was to have the best men in each field appointed as high-salaried chairmen, and to have
about twenty annual prizes of 1100,000 for the best papers of the year. 'As a matter of
fact, any ofthe National Science Foundation bills which were introduced in the 79th and
80th Congresses would perfectly wen serve as a model' The effects would be achieved by
taking the best scientists out of their laboratories, and inducing the others to follow
fashions.
7 See N. Reingold'S review of Or,a";zat'01I for E,tJtUJmi, CO-Ojlerat'on au Dewlopmmt
(D.E.C.D., Paris, 1968), and of E. B. Skolnikoff, Stimee, Te'/moloo .rul Amni,."
Forn", Pol',y (M.I.T. Press, 196'7), and of D. S. Greenbtrg, The Politi" of P",e
Stimee (New American Libnry, New York, 196'7), Isis, S9 (1968), 216-20. Against the
first two, he argues that 'one cannot understand either (social and economic) output or
process without detailed knowledge and understanding of the 'input', which consists of
36 The Varieties ofScientific Experience
it appears that for the discussion of the social problems of science at
a reasonable depth, it can be done. Also, men of administrative
experience have (incorrectly, in my view) dismissed the ethical
component of scientific research as a purely prudential strategy by
scientists with application only to an isolated part of their work. 8
Even more corrosive ofthe ttaditional image of the scientist than
all this explicit discussion are the implicit assumptions that underlie
all debates on 'science policy'. For it is assumed that, when decisions
are made at the highest level on priorities between projects and
entire fields, the work for which support is offered will attract
scientists to it, regardless of their prior interests or views on the
choice. Thus, the bulk of the scientific community can be and is
treated as manpower, which will flow as directed between tasks for
which it has the requisite skills. As we shall see, such policy decisions,
and the resultant flows, are necessary in the conditions of indus-
ttialized science; and the enterprise could not survive if the majority
of competent scientists were extreme individualists who would work
well only on problems of their own choosing. But such manpower-
units cannot be considered as scholars, and any propaganda that
projects the image ofthe typical scientist as an independent searcher,
following his own path in the exploration of Nature, is now worse
than false: it is a bore.
Whether there has been a Fall of Scientific Man over the last
generation is a topic on which any debate will produce more heat
than light. The world ofscience is variegated in the extreme, and our
knowledge of the previous period depends largely on a folk-history
of science which naturally stressed certain aspects of science and
ignored others, and created a simplified, idealized picture out of
disparate and contradictory elements. But we have already seen that
the industrialization of science brings with it certain strong tenden-
specific scientists working in particular scientific contexts'. Greenberg's book he considers
to be 'not about pure science or even about pure scientists', but rather about power,
intematioDal hostility, industrial competition, and regional conflicts; the real focus of
the 'politics of pure science' is not in the scientific establishment but in the State and
industrial agencies that wield real power. In general, R.eingold observes that the one-
sidedness of such books reflects the division and weakness within the history of science,
where 'intemalist' and 'extemalist' approaches, although DO longer tied to hostile
ideologies, make hardly any contact.
• D. S. Greenberg quotes the plain-cpoken Robert M. Hutchins, 'There have been
very few scientific muds. This is becaUle a scientist would be a fool to commit a scientific
mud when he am commit muds every day on his wife, his associates, the president of
the university, and the grocer•••.' (The Polities ofP.t Slimet, pp. Zo-I.)
 Social Problems ofIndustrialiutl Science 37
cies to change, in the realities ofthe social practice ofscience as well
as in its self-awareness. In this chapter I will describe some of these
new problems, and briefly indicate some of their causes. Because
many of the phenomena themselves are ruled out as inconceivable,
let alone inexplicable, in the terms of the traditional philosophies of
science (in particular, the circumstances in which scientific inquiry
proves abortive), it will be necessary first to establish the elements of
a new philosophy of science for their systematic description and
analysis. On that basis we will be able to return to a deeper study of
the social problems of industrialized science, and consider the ways
in which they might possibly be resolved.
The World ofAcademic Science
It will be useful for our purposes to sketch some aspects of the
social activity of science in the period preceding the present one. 9
The institutions and attitudes of science and scientists at the present
time are largely inherited from that period, and the memory of that
era is a point of reference for all analyses of the present except for
those which see science simply as a factor of production. This
period may be considered as having its terminal points in the French
Revolution and in the atomic bomb. The event that marks its
inception is the establishment of the Ecole Polytechnique in Paris,
where scholarship students chosen by ability received a scientific
education from the great masters. Although the school was designed
to provide engineers, and only a small proportion of its graduates
were able to pursue a .scientific career, it provided a model and
inspiration for later scientific education, as distinct from practical
training in medicine or engineering. Within a few decades, first in
Germany and then more gradually in England, science achieved an
institutional base in teaching posts in universities, which themselves
were reformed from their previous stagnation. All the other social
institutions internal to science were either created during this period,
or changed radically from their earlier forms. Generally, dilettantes
and a lay audience were gradually excluded. The work became con-
centrated and specialized, with an apparatus ofjournals and specialist
societies, and an increasing dependence on an occupational base in
teaching and research. The work of science itself progressed at an
9 A most useful collection of papers on the institutional aspects of contemporary
science, seen in the perspective of the recent past, is Sdm&t.rul Sotiet~, eel. N. Kaplan
(Rand McNally, Chicago, 1965).
38 The Varieties ofScientific Experience
accelerating pace; a succession of new fields were opened up for
effective disciplined inquiry, and the increase in factual knowledge
was stupendous. By the later part of the century the centre of
organized research was in the German universities, and there the
system of academic science may be said to have achieved its greatest
moments.
By the end ofthe nineteenth century there were two developments
that in rettospect can be seen to mark the inception of the latter
phase of the cycle of development of academic science as a social
institution existing in a particular, historically conditioned, context.
First, the German universities ceased to expand at their previous
rate; and second, the systematic institutionalized connection of an
important science with industry was established. This·was chemistry,
which had always had a very close relation with practice, but which
in later nineteenth century Germany became the basis for a sophis-
ticated, capital-intensive industry on the pattern which is now
dominant. The example and ideals of academic science, on a model
derived from Germany, remained vigorous well into the current
century. 'Science' has been understood as pure, university-based
science, in spite ofan involvement ofscience in the First World War
so deep that it has been called 'the chemists' war'. Indeed, not until
the Second World War had produced a scientific-technological effort
of a new order of magnitude, culminating in the atomic bomb, did
the interpenetration of industry and science, and the resulting
industtialization ofscience, destroy the claim ofthe 'academic' image
to represent the essential nature of science.
Our description ofthe social practice and self-awareness ofscience
in this past era will be under- four headings: the assurance of a
diffuse social benefit; the ethic of the search for truth; work in the
context of an autonomous community of gentlemen; and the par-
ticularly refined sort of personal property achieved in the work. For
the first two themes, we can do no better than to quote from the
classic pronouncement of Helmholtz:
In fact, men ofscience form, as it were, an organized army labouring on
behalf of the whole nation, and generally under its direction and at its
expense, to augment the stock ofsuch knowledge as may serve to promote
industrial enterprise, to adorn life, to improve political and social relations,
and to further the moral development of individual citizens. After the
immediate practical results of their work we forbear to inquire; that we
leave to the uninstructed. We are convinced that whatever contributes to
 Social Problems ofIndustrialiutl Science 39
the knowledge of the forces of nature or the powers of the human mind
is worth cherishing, and may, in its own due time, bear practical fruit,
very often where we should least have expected it. Who, when Galvani
touched the muscles of a frog with different metals, and noticed their
contraction, could have dreamt that eighty years afterwards, in virtue of
the self-same process, whose earliest manifestations attracted his attention
in his anatomical researches, all Europe would be traversed with wires,
flashing intelligence from Madrid to St. Petersburg with the speed of
lightening? In the hands of Galvani, and at first even in Volta's, electrical
currents were phenomena capable ofexerting only the feeblest forces, and
could not be detected except by the most delicate apparatus. Had they
been neglected, on the ground that the investigation ofthem promised no
immediate practical result, we should now be ignorant of the most
important and most interesting of the links between the various forces of
nature....
Whoever, in the pursuit of science, seeks after immediate practical
utility, may generally rest assured that he will seek in vain. All that science
can achieve is a perfect knowledge and a perfect understanding of the
action of natural and moral forces. Each individual student must be con-
tent to find his reward in rejoicing over new discoveries, as over new
victories of mind over reluctant matter, or in enjoying the aesthetic beauty
of a well-ordered field of knowledge, where the connection and the
filiation of every detail is clear to the mind, he must rest satisfied with the
consciousness that he too has contributed something to the increasing
fund of knowledge on which the dominion of man over all the forces
hostile to intelligence reposes.

Helmholtz then remarked on the likelihood that a scientist will not

receive due recognition for his achievement, but noticed an improve-
ment in recent times. Mter giving many examples of the necessary
interconnections between the various disciplines of natural and
human science, he ended with this exhortation:
In conclusion, I would say, let each of us think of himself, not as a man
seeking to gratify his own thirst for knowledge, or to promote his own
private advantage, or to shine by his own abilities, but rather as a fellow-
labourer in one great common work bearing upon the highest interests of
humanity. Then assuredly we shall not fail of our reward in the approval
of our own consciences and the esteem ofour fellow-citizens. To keep up
these relations between all searchers after truth and all branches of know-
ledge, to animate them all to vigorous co-operation towards their common
end, in the great office of the Universities. Therefore it is necessary that
the four faculties should ever go hand in hand, and in this conviction will
40 The Varieties ofScientific Experience
we strive, so far as in us lies, to press onward to the fulfilment ofour great
mission. 10
The particular form of Helmholtz's statement is conditioned by
the particular traditions in which he, and his audience, parti-
cipated. One was that of the German universities, in which the
cultivation of knowledge was considered as a good in itself, and
also a contribution to the greatness of the nation. The other was
that deriving from the revolution in natural philosophy of the
seventeenth century, with its twin goals of knowledge and power
for mankind.
In other national traditions, the formulation of the themes of
diffuse social benefit and of the search for truth, would be slightly
different; but the underlying similarities are sufficiently great, at
least in those fields which were not deeply involved in technical work,
to justify speaking of a single 'ideology' of academic science; by
'ideology' we mean a definition of reality and a set of values which
serve to guide and to justify the practice of a particular group. A
remarkable feature of academic science is that for at least a century
this ideology could be dominant, even to the point of becoming tacit
common sense, without being modified by any internal strains or by
contradiction with real experience.
This happy situation could persist because of the nature and
social context of the activity of scientific inquiry at that time.
Throughout the nineteenth century, science was a very small-scale
affair. In any given field, there would be only a few small schools of
master and pupils. And opportunities for employment in science
were so few as to offer no atttactive prospects to careerists and time-
servers; those who survived the personal and economic hardships of
initiation into the work were likely to be highly competent and
committed. Hence there was a small set ofoverlapping communities,
whose members were bound together not merely by accident or
convenience, but by the personal ties of shared endeavour and a
common loyalty to an ideal. They had the excitement and gratifi-
cation of opening up one field after another to disciplined scientific
investigation. Because of the nature of their work, and its investi-
gation, its institutional context, and the social basis of recruitment,
most ofthe members ofsuch communities would be gentlemen, who
10 H. von He1moltz, 'On the R.elation of Natunl Science to Science in General', in
Po,.u. LeG'.'s 011 S&inllifit Subj,G'S, 1st series (London and New York, 1893), pp. 1-28;
extracts from pp. 34-8.
 Social Problems ofIndustrialiutl Scimee 41
could be motivated to act by goals more refined than the mere
acquisition of wealth or power.
Because of their size and their membership, these communities of
scientists could manage their affairs with the very minimum of
formality. The tasks of obtaining and administering external funds
for research were few, because of the small size both of the commu-
nity and of the funds involved. The managemel)t of novelty in
science, which inevitably involves conflict, was facilitated in Germany
by the decentralized system of universities, and in England by the
independent status of individual scientists. And the maintenance of
the quality of work done could be accomplished quite informally,
through the close personal ties which linked effective members of a
field with the leadership.
Idealized pictures of this community of academic science can
make it appear as a primitive-communist lay priesthood. It was not
that; not only were there great rivalries and debates, but each
scientist was necessarily involved in the society ofhis time. Each man
had a career to build, and personal interests requiring protection.
This communal function was performed by the system of publica-
tion in journals, where each published paper, certified by a referee or
the editor as good, was a piece of intellectual property of the author.
Others might use it freely, but they were obliged to cite it so that the
value of the borrowed property would be publicly recognized. The
stock ofpublished papers, with their citations in later work by others,
was the scientist's return on the investment of his time and energies.
Although it could not be simply exchanged for cash, it had a
material value of fixed capital investment through its function
as evidence of a man's worth when he was being considered for
advancement.
Of course, the protection of property was not the sole or even the
main function of the system of publication through journals. The
system served primarily as a means to the attainment of the general
goals which defined the community: the advancement of knowledge.
And it did this through the rapid diffusion, to the widest possible
audience, of those results which had been certified for their quality.
In the social conditions of academic science, the two functions
harmonized remarkably well. A man would not want to risk the
disgrace of having offered faulty goods to his colleagues, and so he
would not mind the delay caused by the checking of his results by a
colleague before they were released. By this test, the community was
42 T1Je VMittits of Scientific Ezperimce
protected from the annoyance and waste of attempting to use
unreliable results, and also from the temporary honouring of false
claims to property. And for really urgent news, a private letter to one
or two friends would suffice.
In practice, this form of property is a very subde thing; for
example, the formalized technique of citation cannot always encom-
pass the manifold relations which a result might have to a predeces-
sor. Hence ifthe system is to work effectively in both of its functions,
and not to break down through evasions, there must be a rigid
etiquette for the tteatment of this property. This cannot be enforced
unless the members of the community acquiesce in it; and for this
they must subscribe to a working ethic, itself based on the under-
lying ideology of the activity. To be effective, this needed to cohere
with the daily practice of the scientists, and also with their general
view of the world. Thus the metaphysical belief that atoms of truth
exist and can be discovered in isolated investigation, had its function
in the maintenance of the health of the community of academic
science.
I will discuss these different aspects of the social activity of
science in greater detail later, but at this point one important con-
clusion can be drawn. A healthy and vigorous state of scientific work
is not at all a 'natural' condition, but requires a leadership capable of
providing enlightened direction and imparting morale and commit-
ment to the community. The persistence of excellence in academic
science over several generations in anyone locale, and its diffusion to
new centres, shows that these qualities of leadership could be passed
down from master to pupil. This cannot be by formal precept alone,
but only by everyday practice. It is for this reason that we can be
sure that the ideology of academic science, and its associated ethic,
were really effective, and not merely propaganda and retrospective
folk-history. I I

This idealized account absttaets from the many variations, over

time, between centres, and between fields, which occurred during
the development of academic science. Also, many branches of
science were never in this autonomous and balanced state; chemistry
always had relations with industry, and biology with medicine; and

II I am indebted to Mr. N. Parry for forcing me to realize the necessity of this par-
ticular argument.
 Social Problems ofIntlustrialiutl Science 43
astronomy was always capital-intensive and related to the State. IZ
But the simplified picture on which my analysis is based does have a
certain historical reality: it was the dominant self-consciousness of
academic physical science in Germany in the latter part of the nine-
teenth century, whence it spread to other fields and other nations. 13
As such it is important, for it is such an ideology that provides the
only existing foundation for an ethics of science which is distinct
from that of technological development or commerce. 14 The danger
is that this ideology will be kept in a fossilized state for particular
public-relations functions, while becoming less and less relevant to
the experience of those who live in the world of industrialized
science. Then the inherited ethical principles of scientific work will
12 The basic historical studies of the social and institutional aspects of European
academic science have been made by Joseph Ben-David. His earlier papers are listed in
Scimee flfIIl Society, ed. N. Kaplan (Rand McNally, Chicago, 1965), pp. 581-3. A
pioneering dort to characterize the difference between 'academic science' of the nine-
teenth century and the earlier socW forms, is his 'Scientific Growth: A Sociological
View', Milltl"Wl, 2 (1964), 4SS-T1.
13 In 'The Scientific Worker' (ph.D. thesis, University oC Leeds, 1969), N. D. Ellis
has explored some of the ambiguities in the concept of the 'autonomy' of academic
science. He suggests that the legend oC 'pure science' may be a confIation of two quite
different situations: the German academic, 'pure' with respect to applications, but unless
a full ProCessor, nol autonomous in his choice of problems; and the British amateur, free
to Collow his choice, but not at all averse to involvement in commercial ventures. More
detailed historical studies are required before this insight can be considered as fully
established.
14 Two examples, widely sepanted in context, indicate the great variety of principles
that can be invoked Cor the preservation of the integrity of science. An extreme case is
cited by D. S. Greenberg, Tile Politits ofAmerit_ Sdmee (p. 93), concerning a group of
distinguished biologists in the early twentieth century, who preferred penury Cor them-
selves and for their work, rather than accepting grants from the Carnegie Foundation and
from the University ofChicago which had 'unacceptable' conditions attached. A sttiking
contrast is provided by a comment on a priority dispute between Charles Wheatstone,
F.R.S., and then ProCessor at King's College, London, and an obscure mechanic,
Alexander Bain. In earlier nineteenth-century London, it was not thought wrong for a
man of science to concern himself with pnctical (and profitable) matters; but the
distinction between the two sorts of work was clear to this public. In this case, the
dispute was over a patent, and 80 the reviewer of. book on the controversy could write:
'The ProCessor, in the case before us, has laid aside his gown-he has stepped from his
chair into the shop oC the artisan. His discoveries there are not the treasures presented to
the scientific world, they are riches heaped up Cor his own private emolument ••• when
a philosopher avails himself oC the protection of the Patent Laws, he losts tllSle as a
philosopher, and descends to an equal rank with all others who seek the protection of the
same banner.' The quotation is from a review of J. FinIaisson, A" Atttnmt of SOffIt
Rt1IIMill1Jle Applitflt;oru oftile Elft,"t Fluid 10 lilt UstfiJ bts by M,. AleztllUler B." •.. ,
in the Eltaritfll Mflg.R;ine (October 1843), pp. 139-42. I am indebted to Mr.1l. A. Muir
Cor this reference. An account of Wheatstone's career, with a sympathetic history of his
involvement in another priority dispute, is G. Hubbard, eDDie flnil W1IefltStone.nil tilt
Int~lio" ofllle Eltttrit Telegraph (Roudedge, London, 1965).
44 The Varieties ofScientific Ezperience
become increasingly divorced from reality; and under the pressures
of present conditions, they could not long survive as effective
conttols on action.
The Industrialiution ofProduction in Science
For a comparison of the present age with the one preceding it, we
can best start with the changed technical character of the work of
scientific research; for from this follow the changes in its social
institutions, and social practices. ls The basic difference is a simple
one: research is now capital-intensive. Any significant piece of work
is almost certain to cost far more than an individual scientist can
afford out of his own pocket; it will generally cost much more than
his annual income. Hence he is no longer an independent agent,
free to investigate whatever problem he thinks best. Nor is he likely
to have personal contact with a private patron who will provide for
all his needs. Rather, in order to do any research at all, he must first
apply to the institutions or agencies that distribute funds for this
purpose; and only if one of them considers the project worth the
investment can he proceed.
This change is as radical as that which occurred in the productive
economy when independent artisan producers were displaced by
capital-intensive factory production employing hired labour. The
social consequences of the Industrial Revolution were very deep,
and those of the present change in science, while not comparable in
detail, will be equally so. With his loss ofindependence, the scientist
falls into one of three roles: either an employee, working under the
conttol of a superior; or an individual outworker for investing
agencies, existing on a succession of small grants; or he may be a
contractor, managing a unit or an establishment which produces
research on a large scale by contract with agencies. Of course, he may
have other tasks and responsibilities, as refereeing for journals,
IS I am particularly indebted to Professor D. Humphrey, of Oregon State University,
for helping me develop the ideas of this chapter. I should also say that even before I
bepn to think seriously on these problems, they had been indicated by John Ziman in a
broadcast ta1k, 'Scientists, Gentlemen or Playas?', Tile Listener, 68 (1960), sw-607.
For surveys of the history of the process whereby academic science evolved into industrial
lCience, see D. S. L. Cardwell, Tile Or,fI";UtUm of StinIe, ill £",111l1li (Heinemann,
London, 19S7), and H. and S. Rose, StinIe, fIIUl SodIl;y (Allen Lane, London, 1969).
It is interesting that throughout their historical study, the authors of the latter book
abow impatience with the slow growth of industrialization over the past century; but in
their final chapter they find themselves making a ndical aiticism of the state of affairs
that baa fiDaIly been ad1icvecL
 Social Problems ofIndustrialized Science 4S
advising investment agencies, and so on; but in relation to his means
of production and the decisions which determine his work, his
position must be one of these three.
Along with this differentiation of the positions of individual
scientists, there comes a concentration of the power to make
decisions, and the development of a formal administrative system
for this function. The dispersal of large sums of money, and even
more the decisions between competing demands, are matters which
require proper procedures of information and control. A completely
informal consensus of a large community is not sufficiendy precise
or reliable to be the basis for such work; and the investing agencies
must work from the evaluations and judgements of a group of
advisers. With this concentration of powers of decision and control,
the free market place of scientific results, whose value is established
after they are offered and by an informal consensus, is replaced by an
oligopoly of investing agencies, whose prior decisions determine
what will eventually come on to the market.
This is an inevitable consequence of the costs of research; even if
individual universities were given large block grants out of which to
finance all the research done by their members, the same sorts of
decisions would be necessary, and a similar set of structures would
develop.16 Also, the investing agencies try to maintain some of the
old conditions of consensus, by choosing their advisers from among
the recognized leaders of each field. But the incoming generation of
leaders, and all those following, have built their careers and reputa-
tions under these new conditions, and will be subject to its influences
in ways which I shall soon describe.
The most basic effect of these technical and structural changes is
a tendency to a change in the location ofthe intellectual property ofa
scientist, possession of which is desired for the achievement of his
personal purposes. We recall that under the old system it was
fundamentally the published research report that constituted his
property; on the basis of the informal evaluations of it by his
16 In his broadcast talk on 'Planned Science' in 1948, Michael Polanyi argued mgendy
that 'no committee could forecast the routine progress of science except for the routine
extension of the existing system'. Hence, 'The function of public authorities is not to
plan science, but only to provide opportunities for its pursuit. All that they have to do is
to provide facilities for every good scientist to follow his interests in science.' Polanyi
may have been correct in believing that the Marxist impulse to the planning of science
had evaporated; but the impossibility of carrying out his suggested policy soon made
planning inevitable. See The Logie of LilJerty (Roudedge, London, 1951), pp. 86-90.
46 The Varieties ofScientific Ezperimce
colleagues, he expected appropriate rewards in his career, and the
personal satisfaction produced by public recognition of his work. In
the present situation, the research contract is not merely a pre-
requisite for the future possession of the property embodied in a
published paper; it also brings immediate benefits in itself, in the
way of prestige and possible material conveniences. Moreover, with
the concentration of decision-making power to the investment
agencies and their few advisers in each field, their estimate of a man
is of more practical significance for his career, than that of some
future diffuse consensus. Hence the location ofa successful scientist's
property tends to shift from his published results to his existing
research contracts, and the personal contacts that will ensure their
continuation. With this shift in the location of the scientist's
property, there is a tendency to a corresponding shift in his concep-
tion of a successful career. Especially for someone who enjoys the
role of a contractor, with its incidental benefits, the goals of a
career in science can change from being a series of successful
research projects made possible by a parallel series of adequate
contracts, to being a series of successful research contracts made
possible by a parallel series of adequate projects. I? When this
happens, the man is better described as a 'scientific entrepreneur'
than as a 'scientist'.
Under such conditions of division of labour, concentration of
decision-making power, and tendency to the shifting of goals, it is
impossible for a 'community' to survive in its old form. The mere
expansion in size of every field, speciality, and subspeciality, would
make it difficult in any case. But under the present conditions,
differences in wealth (taken here as a measure of the scientist's
intellectual property) produce such extreme differences in prestige,
power, and material benefit, that we can ttuly speak of classes in a
society ofscience, rather than of more and less eminent colleagues in
a community. Since most scientists engaged on 'pure' research are
employed by a university (or by one of the few State-supported
research laboratories where conditions are equivalent to those at
universities), they will tend to settle at institutions of a prestige
status comparable with their own. The leading men, forming a very
perceptible 'invisible college', will congregate at the great univer-
.7 This distinction is made implicidy by D. S. Greenberg when he contrasts 'grants-
manship', the (ethically neutral) practice ofextracting funds for research, and 'chiselling'
(Tite Politi&s of Alllerita Stim&'t pp. 351-3).
 Social Problems ofIndustrialized Science 47
sities, enjoying favourable conditions of employment there, and with
their intimate connections with the investing agencies (a prerequisite
for the existence of such a group) they will pursue the researches
they please in comfort. And since in most countries the high-
prestige universities cluster in a special geographical region, those
who are left, or cast, out of the charmed circle of the successful men
may be isolated in every way. Deprived of the personal contacts
whereby one keeps up with new work, and whereby one makes an
impression on those who advise the investing agencies, they are left
to do derivative or second-class research on less generous grants, or
none at all.

Adulteration ofthe Products ofResearch

We can now consider some of the changes in social practice
consequent on the changes in production in science; and in so doing,
identify and explain some recognized social problems. The first of
these has been the subject of anxious discussion for some years: the
'information crisis' .18 At first this appears to be a purely quantitative
problem; the number of journals in each field is already so large, and
is increasing so rapidly, that a scientist cannot 'keep up' with the
literature without great difficulty.19 Worse, because of the difficulty
of finding any particular item, even with the existing abstracting
services, scientists frequendy suffer the waste and disappointment of
duplicating research already done elsewhere. Several schemes have
been developed for more effective information retrieval systems; but
it has to come to be seen that the complexity of the contents of a
given research report makes purely mechanical identification less
valuable than was at first thought. ~o
On closer examination~ it appears that the problem is not a purely
quantitative one. For along with the expansion of the traditional
II An illuminating survey of problems of communication in science, including some
striking examples of blunders in research that appear in print, is given by H. V. Wyatt
in his essay 'Communication in Obscurity' in Tile Uu of Bio¥1I1 Lit.",,,,, eels.
R. T. Bottle and H. V. Wyatt (Butterworth's, London, 1966).
19 See Bendey Glass, 'Information Crisis in Biology', BuIleti" oftile AltnIIie Seientistl,
18 (October 1962), 6-12. He reports a survey of fifty scientists, in both physical and
biological sciences, on their ~ of the literature; the conclusions suggest a 'growing
isolation in which much scientific research is carried on'. He also deplores the decline of
genuine review articles, as opposed to mere annotated bibliographies; this can be
explained by their not counting as 'research' by the author.
zo See P. Cranefield, 'Retrieving the Irretrievable; or the Editor, the Author and the
Machine', BldJeti" ofthe MeJi&1I1 Li"'Il,y Assotilltitm, ss (1967), 12C)-34-
48 The Varieties ofScientific Ezperimce
channels of publication, there has been a rapid development of less
formal channels. Through mailing-list distributions, scientists will
circulate not only reprints, but also preprints, duplicated preliminary
research reports, conference abstracts, and informal 'newsletters' of
people, events, and results. These other types of publication are not
merely supplements to the official channel of communication,
making results available more quickly than can the printed journals.
Rather, they are complementary to that channel; and their function
is to provide publication ofa sort which is not subject to the hazards,
as well as the delays, of the scrutiny of referees. For the same sort of
function is performed by the raw collections of conference papers,
published in hard covers and sold on the market, but appropriately
called 'non-books'. The fact that this channel is a dilution of the
other is recognized in discussions of the problem of' deciding
whether such materials can be cited in a paper published traditionally.
For the materials so cited should not only be public, they should
also be of tested quality; otherwise a research report which has been
certified by publication will be dependent in part on a component
which is uncertified. Of course, there is no easy answer to this
problem; for whatever the criterion of distinction between orthodox
and unorthodox publication, some clever fellow will find a technique
for straddling it. ~I
Thus the problems created by this diluted channel are far greater
than the creation of additional headaches for those who do the work
of abstracting and information retrieval. For they show that a large
number of scientists are quite happy to evade the traditional rules
for publication, and to be seen doing so. This phenomenon can
easily be explained in terms of the shift in the location of intellectual
property. Through the adulterate<fchannels, one can hope that one's
results, dressed up in their most attractive form, will catch the eye of
someone important, where another routine paper in a crowded
al In 1960, the editor of PilYsUIJI RnJin, Letters took a firm stand against initial
publication of results in the daily press, remarking that 'Scientific discoveries are not the
proper subject for newspaper scoops', and warning that previously released results
would not be accepted by his journal. See D. S. Greenberg, Tile Poli,it$ ofAmerica"
Scimel, p. 74-) A decade later, and in a field in which pressures might well be Sb'Onger,
the barrier was surrendered. Dr. Morris Fishbein, the eminent leader of the American
Medical Association, wrote: 'The distinguished editors ofthe clinical journals would do
a disservice to the medical profession by demanding that medical investigators withhold
information about their observations and conclusions until sufficient time has passed to
allow the medical journals ofrecord to publish their work in its totality. Nor should they
expect investigators to heed such directives.' (Q!toted in Sdmel, 167 (1970), 148.)
 Social Problems ofIndustrialized Scimce 49
journal, or a bare tide in an abstract, would be passed over without
notice. But to engage in this practice, a man must be willing to put
untested products on offer to his community, knowing that some of
the work might be substandard. 22
That there is no shortage of such people, becomes clear when we
discover that one of the most serious aspects of the 'information
crisis' is a phenomenon known as 'poindess publication'. But since
every aet of publication has some purpose, I shall refer to it as
'shoddy science'. For it appears that the majority of journals in many
fields are full of papers which are never cited by an author other than
their own, and which, on examination, are seen to be utterly dull or
just bad. 23 Now, the existence of bad scientific research is something
of which every scientist (or at least every good one) must be aware
by his experience of attempting to use other published results. But
hitherto it has not been a topic for polite discussion in print; it is
mentioned neither in the philosophy of science nor in the formal
teaching in science. Research students learn of it, frequendy by
very painful experience, as something of a dirty secret. Indeed, not
very long ago an elder statesman of science could assert that such a
thing is nonexistent.

Would it be too much to say that in the natural sciences today the given
social environment has made it very easy for even an emotionally unstable
person to be exact and impartial in his laboratory? The traditions he
inherits, his instruments, the high degree of specialization, the crowd of
witnesses that surrounds him, so to speak (if he publishes his results)-
these all exert pressures that make impartiality on matters of his science
almost automatic. Let him deviate from the rigorous role of impartial
experimenter or observer at his peril; he knows all too well what a fool
So-and-so made of himself by blindly sticking to a set of observations

22 The American Psychological Association baa recendy found itself in mnfliet over a
project designed to provide an institutional machinery for the publication of untested
materials. An 'early dissemination' scheme, which forms one part of a mmprehensive
plan for improving communications, would distribute manusaipts to special-interest
subsaiption lists shortly after submission. This has been aitic:ized as 'a vast sewer
carrying garbage from one scientist to another'. Disquiet over the very principles of the
system, which was designed to mpc with a genuine information crisis in the subject, is
mixed with mncern that the bureaucracy necessary for running it will dominate the
Association. See P. M. Boffey, 'Psychology: Apprehension over a New Communications
System', Scimet, 167 (1970), 1228-30.
23 D.]. de S. Price, Bi, Scime" Link Same" reports a paper by D. J. Urquhart on
loans of journals in the stock ofthe Science Libnry in London; about half were not used
at all, and a quarter only once in the year of survey (pP. 75-6).
SO The Varieties ofScientific Ezperitnce
or a theory now clearly recognized to be in error. But once he closes the
laboratory door behind him, he can indulge his fancy all he pleases..•. ~4
The most remarkable thing about this argument is that it assumes no
special virtue on the part of the scientist; but it does claim that the
system of quality conttol in science is perfect. The claim may have
been plausible at some time in the past, but it is no longer so.
Shoddy work exists, and in large quantity. References to it can be
gleaned from published discussions of the state of particular fields.
And it is a truly pathological symptom of the social condition of
industrialized science.
For a paper to be published, it is sufficient that the author, an
editor, and a publisher all find some purpose served by its publica-
tion. From the side of the publisher, it is a matter of economics:
given the guaranteed library subscriptions and the economics of
journal publication, it is possible to make a profit even on a obscure
journal.~5 The editor may receive an honorarium from the publisher,
and will certainly derive prestige at his own university by virtue of
his position. The author may need to have another tide in his record,
as a demonsttation (for his employer) of his continued competence
in research, or as another point to be included in his aggregate score
of publications when he applies for a grant from a large and imper-
sonal investing agency.
Of course, the publication of shoddy work would exclude a man
from membership in a community devoted to the advancement of
science; hence those who do this are either in no community at all,
or in one with different goals. In such a community, the traditional
ethic of the disciplined search for truth is either forgotten, or is a
sick joke. 26 If the publication of ~oddy science were restricted to
the dim and obscure men in remote provinces, and their local
journals, then the situation could be explained partly at least by the
natural differentiation by quality of scientists. But the participation
in this abuse by men of high prestige requires another explanation.
24 James B. Conant, 0. UtUlerstlUlllillK StinIe, (Yale University Press, New York,
1947; Mentor Boob, 1951), p. 23 of Mentor edition.
as See 'How Many More New Journals?', N""", 186 (2 Aprillg(0), 18-19-
:a6 For a penetrating analysis of the problem of shoddy science, see Howard A.
Mayerho~ 'Useless Science', BtJleti" ofllte AltnIIie Scimtists, 17 (March 1961), 92-4-
This article is a reprint of a speech originally given in 1954 by the author, a former
editor of StinIe,. He describes the various pressures on scientists to accumulate pub-
Iication-points, and gives useful case studies on how iDsignificant research can be made
fruitful and multiplied.
 Social Problems ofIndustrialized Scimce 5.1
The Penetration ofScimce by Industry
It is well known that science, even the most pure science, is
important for the work of industry. Any firm engaged in modem
industrial production, and by extension an advanced industrial
nation, secures its future existence by the work done in the 'research
and development' laboratories. Such work does not merely use
scientific results; in its more sophisticated forms, it is continuous
with scientific inquiry. Problems and discoveries thrown up in such
technological inquiry can stimulate important scientific research;
and conversely, any scientific result, even one deriving from a
problem which was investigated with no thought ofapplication, may
find a use in industry. All this should be familiar; and in this
unification ofscience and technology lies the hope for the realization
of the age-old dream of material plenty for all of mankind.
What is less familiar is that industry has, in many ways, made its
own penetration into science. The industrialization of scientific
research is one manifestation of this. By itself it involves an increase
in scale, and in formal organization; but we can be sure that the vast
expense ofindustrialized science would not have been incurred by the
State unless some tangible benefit through a close association of the
two sides was expected. How close the association can be was
indicated in a report on the ethical problems of physicists in America
a few years ago."? There it was found a man could be simultaneously
filling nine roles: at his university, to be teacher, administrator, and
research scientist; with various State agencies, to be a contractor for
research, an assessor of research proposals, an official adviser on
existing projects, and a private consultant on particular technical
problems; and with commercial industry, to be a private consultant
to firms, as well as a businessman manufacturing equipment of his
own invention. It is hardly surprising that little financial abuses
should occasionally occur, nor that 'conflict of interest' should be a
matter of concern. What might be asked is whether such a man could
change his attitudes and values every time he picks up a different
piece of correspondence, or whether he operates in a world where all
distinctions are blurred, and it's all business.
The penetration of science by industry proceeds by yet another
27 See 'Hazards of Sponsored Research', Pltysics Totllly, 18 (March 1965), 98, 100;
a repon of a statement issued the previous December joindy by the American Council
of Education and the Council of the American Association of University Professors on
the prevention of conflicts of interest in govemment-sponsored research at universities.
S2 The Varieties ofScientific Experience
path, created by an inevitable ambiguity in the specification of every
research project. This is in the possible function of the result to be
achieved. Even if the scientist is personally interested only in its
significance for advancing the field, he can well imagine a possible
relevance of the result to some sophisticated and unsolved technical
problem. An emphasis on this relevance will help the project to be
considered as 'mission-oriented research', which is naturally better
endowed with funds than the useless sort. Now, it is possible for the
scientist or the investing agency to keep the different aspects of a
project, and the different values they represent, in separate compart-
ments, considering it accidental and unsought good luck that
interest in the problem should be so diversified as to bring it
financial support. But such a state of affairs cannot persist indefi-
nitely for a whole group of scientists. The natural tendency, to
ensure good relations and continued support, is to give serious
attention to those whose interest is essential and for the research to
shift in their direction.
Of course, the community of science has always had a very
difficult problem in justifying itself in terms comprehensible to the
lay world; and in its generalized claims to be the basis for progress in
industry it was to some extent playing a gentle confidence game with
its public. But it did not matter too much, for the resources devoted
to science were very small, and of them only a fraction came from
the public purse. 28 Also, the claims on behalf of the practical
importance of science were necessarily diffuse or retrospective; for
it was only rarely until the end of the nineteenth century that any-
thing like systematic scientific inquiry could be successfully applied
to the solution of industrial problems. But in the present period the
sums are significant, and the claims are true enough for those who
advance the cash to feel entitled to see some return. Hence the indi-
21 On occasion, the mnflict of values between excessively pure-minded scientists and
• particularly vulgar governmental patton muld yield disasters. See G. D. Nash, 'The
Conflict Between Pure and Applied Science in Nmeteenth-Century Public Policy: the
California State Geological Survey, 1~1874', Isis, 54 (1963),217-28. Norman Kaplan
uses the term 'bounty' to desaibe the grants formerly made by European governments
to individual scientists for the pursuit of their research; these would be on a small scale,
and recognized as conttibuting to the research work ofa university for the enhancement
of its, and the nation's prestige. See N. Kaplan, 'The Western European Scientific
Establishment in Trausition', AfI'Ieri&tm Belulvi.,,,l Seientist (December 1962), 17-21,
reprinted in Stinlee fIIUl Soeiety, ed. N. Kaplan, pp. 352-64- It should be observed that a
very important part of this 'bounty', at least between the two wars, came from American
foundations, notably the Rockefeller.
 Social Problems ofIndustrialized Science S3
vidual scientist, and even more .those leaders of the scientific
community who plead for public funds for particular projects as well
as for general purposes, must be able to talk the language ofeconomic
(or military) benefit at least as well as that of the search for know-
ledge.
Influences from Runaway Technology
The relations of science with industry will not be uniformly close
in all fields; the connections will be strongest with the most modem,
rapidly developing technologies, where innovation depends entirely
on large-scale, sophisticated 'research and development'. It is these
areas, such as aerospace, eleettonics, and parts ofbiological engineer-
ing, where the pace of development is so rapid, and the ecological
and social effects so unpredictable and dangerous, that have been the
focus of public concern in the menace of 'science' to humanity.
Those who take the decisions to plunge into ever greater 'progress'
in this work are not aftlieted by any special wickedness or even
irresponsibility. They are merely continuing the attitudes and
practices inherited from the industry of the past, which might be
called 'myopic engineering'. Provided that a particular development
was technically viable and not at risk of penalties under the law, then
so long as it seemed likely to make a profit, it would be adopted with
no further thought of its consequences. Hitherto, the effects of such
a policy, however disastrous, were localized to the region where they
were put into practice; thus the rural South of England knew little,
and generally cared less, ofwhat was being done to the Midlands and
North by the Industrial Revolution. And it is undeniable that the
generally short-sighted and ruthless men who created the industry
of the nineteenth century laid the material foundations for the
prosperity of the present.
But the engine of innovation and production which we now
possess is qualitatively different in many respects. Its effects are so
pervasive that there is no place to hide from them. Also, the work of
innovation in the advanced technologies is now a large industry in
its own right. Its projects have some special features, which make
them very different from the work of inventors and scientific
consultants in the past. First, the investigation of any technical
problem ofdevelopment requires the prior commitment ofenormous
resources, both in funds and skilled manpower. Also, such problems
are necessarily speculative, in several ways. It cannot be guaranteed
S4 The Varieties ofScientific Experience
in advance that the research will produce a device which works at
all. Moreover, even if it works, there is the risk that during its years
of development, either a change in the technical or commercial
context, or the appearance of a better suited product from a com-
petitor, will deprive it of its market. Perhaps most significant of all,
the concept of 'profit' has been transformed. For much of the
sophisticated technological work is done for the State, for use outside
the market sector, as for war. Although the prospect of foreign
sales of a device are part of the calculation of 'benefit', the basic
component of benefit is assessed through a scientific study of its
potential uses. Even where a key industry is nominally in the private
sector, the State will take responsibility for its continued prosperity,
through research contracts, guaranteed purchases, and other tech-
niques. Thus a particular innovation may be recognized as risky
from the technical point of view, dubious from the commercial point
of view, of very slight use to anyone at all, even the State, and a
potentially serious nuisance to the public and source of legal and
political difficulties, and yet still receive enormous sums from the
State because of its contribution to national prestige and its impor-
tance for maintaining employment and morale in a key industry. The
Anglo-French supersonic transport is a perfect case in point of this
phenomenon. At the extreme, national prestige may become so
involved in a project that all considerations ofcost, benefit, and profit
(except, of course, to the private firms doing research and manufac-
ture) are cast to the winds, and a glamorous technical project, such
as the moon-race, absorbs resources on a gargantuan scale, all in the
name of 'science'.
Although this new industry of'R. and D.' employs many scientists
(indeed, the bulk of graduates in science and technology go there
rather than into teaching or university research), its working ethics
are descended from industry, private and state-supported, rather
than from academic science. In America, the enormous defence and
aerospace industries carry on in the time-honoured American
ttadition of 'boondoggling' on Government funds; the most effective
path to super-profits being to keep the relevant Government
agencies for cost-accounting and quality-control either remote, or
weak, or complaisant. 29
at The first extended discussion of this problem is in H. L. Nieburg, /" the Nfl""
ofSlime, (~gle Boob, Chicago, 1966). A summary appears as 'R and D in the
Contract State: Throwing away the Yardstick', BulI,ti" of lhe Atomit Sti",tists, 22
 Social Problems ofIndustrialized Science SS
Thus we can speak of this new technology as 'runaway' in several
respects. In calculating cost and benefit, it ignores all those costs of
a project for which it cannot legally be called to account: in par-
ticular, the degradation of the natural and human environment.
Since the combined effect of the present and future technological
developments is likely to be catastrophic, this rush onwards can truly
be considered as out of control. And in its internal workings, the
absence of that traditional discipline, crude and frequendy distorted
but in the last resort effective, of the test of a commercial market,
makes the category of 'profit' an artificial one, to be determined by
the judgement of men in State agencies, in C<H>peration with the
promoters themselves.
It is in the borderland between science and this sort oftechnology
that we find some significant pathological phenomena. The first
occurs when a contractor (individual or institutional) develops a
really big enterprise, which is most likely to be on some mission-
oriented research in a field where money is plentiful and not too
many questions are asked. 30 There then develops a research business,
making its profit by the production of results in the fulfilment of
contracts. The director of such an establishment is then truly an
entrepreneur, who juggles with a portfolio of contracts, prospective,
existing, extendable, renewable or convertible, from various offices
in one or several agencies. The business is precarious, of course, for
his only capital is in his friendly contacts with those who decide on
the allocation of funds. In such a research factory, conditions are not
usually conducive to the slow, painstaking, and self-critical work
which is necessary for the production of really good scientific
results. Hence much, most, or even all the work can be shoddy; but
(March 1966), 20-4- A recent, more general survey of American military procurement
and spending is w. Proxmire (Senator), Report from WastelatUl (Praeger, New yom,
1970). For the repon of the Lang Committee on the English Bloodhound missiles
contract see Tite Times, 10 February 1965.
30 D. S. Greenberg has invented a prototype entrepreneurial scientist, 'Dr. Grant
Swinger'. In 'Grant Swinger: Reflections on Six Years of Progress', Scimel, 154
(16 December 1966), 1424-5, Dr. Swinger describes the exploits oCthe 'National Animal
Speech Agency', which chad its origins in the President's challenge to the nation "to
teach an animal to speak in this decade"'. Previously, Dr. Swinger had advocated a new
model ofhigh-energy particle accelerator which would extend from Palo Alto, California
to Cambridge, Massachusetts. It would be shaped so as to pick up maximum local
political support. An alternative was a vertical accelerator at the unique intersection of
four States, in the West; the design yielded. very attractive beY/dollar/vote analysis.
See '1965: Herewith, a Conversation with the Mythical Grant Swinger, Head of Break-
through Institute', Scimee, 151 (I January 1965), 29-30.
 S6 The Varieties ofScientific Ezperitnce
the entrepreneur does not operate in the traditional market of
independent artisan producers who evaluate work by consensus. So
long as he can keep his contacts happy, or at least believing that they
personally have more to lose by exposing themselves through the
cancellation or non-renewal of contracts than by allowing them to
continue, his business will flourish. 31
It is in such circumstances that a man of high prestige and real
talent will produce a stream of shoddy work. Too busy to do any
thinking himself, and yet requiring a steady stream of publications
as a proof ofhis continued competence, he will toss off pieces, either
alone or with associates, which will produce a list of titles of the
necessary length. Although large-scale science is more exposed to the
risk of invasion by enttepreneurs, size is not the determining factor.
Whenever a research contractor, however modest his plant, sets the
goals of his establishment to be the renewal and extension of con-
tracts rather than the achievement of worthwhile results, he is an
entrepreneur.
Even when scientific work of good quality is being done, the style
of runaway technology can infect a field; the old, diffuse ideal of
material benefit gives way to something more sharply defined and
intoxicating: the possibility of the creation of new technical powers.
The patent dangers of some of these powers, in the present state of
civilization, have been brushed aside as of little consequence, or as
the responsibility of someone else. It is so many generations since
people in our civilization believed that there are 'secrets too powerful
to be revealed', that a scientist of our age cannot conceive himself as
being in the position of the sorcerer; and yet he is. Thus 'reckless
science', as a special product ofthe technical and social conditions of
scientific inquiry in our time, must be identified and controlled, for
the safety of humanity in the long run and for the preservation of
science in the short run. 32
31 For a satirical but none the less penetrating acmunt of the techniques of entrepre-
neurial science, see H. Miner, 'ResearchmaDship: the Feedback of Expertise', Hurna"
Or,aiutitm, 19 (1960), 1-3. The reference comes from J. Barzun, S&ien&e the Glorious
E",en';",.", (Seeker Ie Warburg, London, 1964), where this and many other targets are
put under heavy fire.
32 The leading candidate for the status of 'reckless science' at the time of writing is
mo1ecuJar genetics, which can enable the controlled manipulation of human genetic
material. See S. L. Luria, 'Modern Biology: a Terrifying Power', The Nation (20
October 1968), 406-9. One of Luria's most promising students, James Shapiro, has
taken the message and left science. He pve three reasons, of which the first is his fear
that his rauIts will be 'put to evil UICI by the men who mntrol science'. See J. K. Glass-
man, 'Harvard Genetics R.esearcher quits Science for Politics', S&ien&" 16,(1970 ),963-4-
 Social Probltms ofIndustrialized Science 57
Finally, the demands of military technology in particular provide
opportunities for employment ofscientists on research projects whose
intended application lies beyond the pale of civilized practice and
morality. The weapons called 'ABC'-atomic, biological, and
chemical-are each, in their own ways, morally tainted. Research and
development of such weapons can be plausibly justified in terms of
defence and deterrence; but the .experience of the scientists on the
original atomic bomb project shows that once the weapon is available,
the tender consciences of the scientists who created it will not have
much influence on the decisions on its use.
These four abuses, shoddy science, entrepreneurial science,
reckless science, an'd dirty science, are distinct in their natures, but
there will be tendencies for them to overlap in practice. 33 Also, each
of them arises from conditions inherent to the situation of contem-
porary science; and there is no clear line of demarcation, in the
results of the work or in the attitudes of the scientist, whereby one
can condemn one man and exonerate the next. But because they
are more closely related to the demands of modern technology, and
sometimes more easily popularized as exciting than traditional
research, they will tend to attract a lion's share ofthe available funds,
thereby providing the most attractive career prospects for recruits,
and drawing into their ambit those who could not otherwise carry
on their research. This effect can be seen in the United States, where
the total budget for 'science' is enormous, but where all save a small
fraction is allocated to military R. and D. and the space-race. Even
there, it can be argued that Congress would never allocate more than
it does for the direct support of the esoteric and peculiar activity of
pure research; but the result is that the 'scientist' is seen as costing
a lot of money to the taxpayers, and if he wants to use some of it, he
is under pressure to make his accommodation with those who con-
trol these branches of runaway technology.

Problems ofMorale, and ofMorals

Up to now, our discussion has been of problems which could be
shown to arise naturally out of the changed technical and social
conditions of science, and whose existence can be established by
representative instances. The problems we shall now discuss are
33 A case allegedly illustrating such tendencies is cited in s. M. Hersh's CIIemi&,,1
au B;olo,;cal Warfue; America's Hidden ArsnuJl (MacGibbon & Kee, London, 1968),
pp. 141-2. It was from the station mentioned there that a cloud ofDerve gas killed six
thousand sheep in 1969.
S8 The Varieties ofScientific Ezperitnce
more s\1bde, although equally important; and for evidence I shall
need to use certain recognizable symptoms rather than hard cases.
The need for good morale is never mentioned in general discus-
sions of science directed to a lay audience; and this is evidence that
hitherto its presence could be taken for granted. For doing good
scientific work is strenuous and demanding, and the quality of the
work done in any field of science is dependent, to a great extent, on
the integrity and commitment of the community of scientists
involved. In other spheres of social activity where the success of the
enterprise cannot be assured by the imposition of systems of disci-
pline and conttol, the factor of morale is given due recognition.
Military organizations are the outstanding case in point; but volun-
tary associations of any sort are known to require good morale; and
nearer home to science, the performance of students at a university
is also seen to depend on their morale. The maintenance of morale is
a most subde task; it is easiest in a voluntary community of equals
doing successful work; and where there are gradations of status and
power, the leaders must show themselves as standing for the whole
group, in its problems, its interests, and its difficulties. 34
Whether there has been a decline in morale affecting all of science
over the last generation is impossible to say. Even if one were to
conduct a social survey of a large sample of scientists, the evidence
for the past would have to come from the personal memories of older
men, and these are not reliable as historical testimony. But there is
one recognized phenomenon, which is probably a symptom of
declining morale, and certainly a cause of its further decline. This is
the disappointment in the recruitment to science, measured by the
number of students electing to study science at university and also
continuing into research. The cause, or causes, of this phenomenon
are not at all understood. It may simply be that the proportion of
young people who, in the present conditions of our culture and
educational system, have the inclination and talent to pursue science
as a career, is strictly limited; and that merely making more places
14 An excellent analysis ofthe importance of good morale for scientific work is given in
BiDe1lemistry, 'Moleeu1M Biolou' IDUl Biololie"l Stinlees (The Biochemical Society,
London, 1969), 26-,. The argument is extended from science to the universities, as a
source of. necessary idealism in modern society, and for which excellence, and 'in the
last raort, excellence of morale' is of the utmost importance. This emphasis on morale
in science and in universities can be explained by the circumstances of the production of
the report as an answer to the 'Kendrew Report', Report of tile Woriill' GrOllP 011
Moleeu1M BioltJo, Cmnd. 3675 (H.M.S.O. London, 1968), whose recommendations
could be interpreted as favouring a contraction of research to a few centres.
 Social Problems ofIndustrial;zed Science 59
available at universities does not, in the short run, increase the
supply of worthwhile candidates. It may also be that many young
people of the classes which provide the recruitment to universities
do not, as did their parents, have a keen awareness ofa need to obtain
a qualification leading to a secure job. So if their personal interests
lie in the arts or in society, they will be less inclined to sacrifice them
during their university studies.
Another possible factor is the decline in the quality of science·
teaching in the schools. A bright but uncommitted pupil can easily
be put off such an abstract and demanding discipline if he never
encounters a teacher who by his personal example makes it an
exciting challenge. And through failing to ensure that well-qualified
and well-motivated scientists returned to school-teaching the com-
munity of science was, for two decades after the Second World War,
eating its seed-corn. Even if the supply of science teachers now
improves, there is no guarantee that it will soon yield to science a
crop ofrecruits ofthe desired quality. Pleasant classroom experiences
are not the only factor that influences the choice of the bright pupils,
to say nothing of those who are really gifted.
It is possible (one can say no more in the absence of very sophis-
ticated social investigations) that part of the falling-off of recruit-
ment to science is a result of the changed image of science which is
projected to the public at large and school-children in particular.
Throughout the nineteenth century many of the great men of
science shared their enthusiasm with a wide public, through books
and lectures. What they conveyed was the sense of excitement of the
individual search for truth through fascinating new discoveries. The
endeavour was one which offered the hope of fame, through work
which was innocent, enjoyable, and ultimately beneficial to mankind.
In the present age, there is no lack of such propaganda; but its
message does not come through in such a clear and simple fashion.
First, the dangers inherent in science have been common knowledge
ever since the menace of nuclear weapons became a matter of public
concern. Also, working in science is now recognized as a career, one
among the many open to a person of the requisite ability, rather than
a vocation for dedicated individuals. For the great expansion of
university places in science is not seen as resulting from an enligh-
tened desire for more pure knowledge, but as a response to a need
for more manpower ofa particular sort. Thus, to put the matter very
simply, the effective image of scientific work may have changed from
60 The Varieties ofScientific Experience
being pioneering explorations in the philosophy of nature, to the
service of technology, either commercial and vulgar, or military and
sinister. And what sort of ambition is it, to be a unit of scientific
manpower?
This is not a fair representation of the state of science today; but
here we are not discussing a complex and subde social reality, but
rather the simplified and distorted picture of it which reaches a lay
public. And it may be significant evidence that the field which has
suffered most from a fall in recruitment is the one whose public
image has undergone the most drastic transformation. 35 Through
the later nineteenth century and well into the twentieth, physics was
the queen of the natural sciences. Its achievements were the most
impressive, it stayed the closest to the philosophy of nature in name
and in spirit, and it was accepted, by philosophers and laymen alike,
as the paradigm ofscience, setting not only the standard ofexcellence
but even the popular conception of what every science should try
to be like. It was a tragedy for many eminent physicists, and also for
their field, that this most aristocratic and philosophical of studies
should have been the one which created the runaway military tech-
nology of nuclear weapons. By this work, physics was tarnished
beyond repair; and through its status as the paradigm field, its moral
fall affected the rest of science as well.
In its present technical and social situation, the world ofscience is
particularly vulnerable to a decline in recruitment. For much of the
research done at universities depends for its accomplishment on a
supply of the poorly-paid skilled manpower of postgraduates. There
is a rapid turnover of this labour, and as it declines in quantity and
quality, each scientist will see his research suffering a similar decline,
and with it his career prospects.36 As we have seen, neither the
35 The 'swing &om physics' is not confined to Britain; the same phenomenon has been
observed in America, and the imaginative new secondary-school syllabuses designed by
leading research physicists have certainly not had a positive effect on the problem. See
D. S. Greenberg, Tile Politits ofAment." Stimee, p. 67.
Statistics for the decline of physics enrolments in American high schools are provided
by F. G. Watson, 'Why do we need more physics courses ?', Tile Ph,sits TeM1Ier
(May 196'7), pp. 212-14- He gives percentages of twelfth grade pupils taking physics, u
follows: 1948-9, 26 per cent; 1958-9, 25 per cent; 1964-5, 20 per cent. This drop of
about 20 per cent in eight years may be partly the result ofa higher proportion ofthe age-
group being enrolled in the twelfth grade, either by a policy ofautomatic advancement in
grade regardless of performance, or by a raising of the school-leaving age in the back-
ward States.
36 Individual impressions ofa decline in quality ofstudents are very suspect, but there
are examples where objective criteria can be cited. Thus, Bart J. Bok (letter to Sdmte,
 Social Problems ofIndustrialized Science 61
technical conditions of research, nor the social experience of scien-
tists over the last two decades, make it easy for them to adjust to the
situation by simple belt-tightening. Without the prospect of attrac-
ting interesting and useful research students, even university
teaching in science loses its attraction and much of its present
function. Hence, whatever the causes of the fall in recruitment, its
effects are plain; and as the decline in morale within science becomes
known outside, it may well contribute to a cycle starting with a
further disappointment in recruitment. Where this process can end,
no one can say; and it is possible that a change in the economic and
social situation of potential recruits will bring them back again in
satisfactory numbers; but whether even this would provide the
leaven of gifted scientists in the mass of technological manpower,
only time can tell.
It is in physics again that we find another symptom of a loss of
morale: the sense that there are no more challenging problems
which can be solved. 37 In the period before the Second World War,
the most exciting part of physics was the study of the atom; and as a
result of the technological success with atomic and nuclear weapons,
physicists were in a position to apply ever larger energies, at ever
larger cost, to the penetration ofmatter in the small. But the increase
could not continue indefinitely, and recendy a halt was called, first
because of the enormously greater expense of each new accelerator,
and also because of growing doubts about whether even such
monsters would be adequate to their intended functions. Nuclear
physics now finds itself at a dinosaur stage; unable to evolve further,
it awaits extinction unless some happy accident rescues it. Of course,
there were prophecies of doom in the leading sciences at the end of
both the eighteenth and the nineteenth centuries, which turned out
to be false. But in each case, the new line of progress came from an
obscure corner of the field, rendering much of the earlier work, and
the eminent men who did it, obsolete.
The sudden discovery of an upper limit to financial support has
not been restricted to physics. All over the world, governments are
having another look at the budgets for research, and at the very least,
154 (I g66), 590-"2) describes aspects of a general relaxation of the pace ofgraduate study
in astronomy, and the lessening of the demands made on students (in nte of progress,
and conttibution to discussions and teaching).
37 H. S. Lipson, F.R.S., in 'Where is Physics Going', AtlWl1l&e, NO.4 (1g68), 17-21
(U.M.I.S.T.), speculates on whether physics u a school subject may share the fate of
Latin.
62 The Varieties ofScientific Experience
decelerating the expansion of support for pure science. It matters
not that the total budget for genuine science is only a small propor-
tion of the sums allocated for technology of various sorts; when cuts
have to be made, they fall most heavily on those with the least attrac-
tive case to present. Ifthe resulting contraction were to be properly
handled, it could have beneficial effects; for science would be
healthier without the mass of dull and shoddy work which is sup-
ported when money is easy. But the danger is that those who main-
tain their contacts with the investing agencies will manage to survive,
while those on the outside, however worthwhile their work, will go to
the wall. The problems of decision on the allocation ofresources are
particularly painful and difficult when there is not nearly enough to
go around; for a fair distribution, exceptional qualities of leadership
are required. Only then could science survive this contraction with-
out a very serious loss of morale.
These problems ofmorale pose a threat to the continued existence
of scientific inquiry as we know it, which is as serious as those
arising directly from technical and social conditions of industrial-
ization. For when morale is low, the work which is done is at the
minimum level of acceptance; and decisions and evaluations which
are reached through consensus will tend to be those which involve
the least risk and the least work, all around. Since the leadership of
science exists and operates only by consensus, the quality of that
leadership, to which the quality ofscientific work is very sensitive, is
at risk. The situation is further aggravated by the fact that for most
ofthese problems, the inherited working ethic ofscience is irrelevant,
just as its ideological basis is obsolete. Hence in many respects,
individual scientists, and even more the leaders ofscience, encounter
problems directly involving morality.
As we have discussed, science has traditionally been porttayed as
an activity whose morality could be nothing but the best. The search
for truth is innocent and ennobling; and the eventual benefits to
mankind through the advance of knowledge and power, further
secure the moral status of science. The very idea of a scientist being
a thief, a swindler, or a man who offers his opinions for sale, is near
to being a contradiction in terms.
The contrast between the working ethics of science, and that of
politics, or business, is so strong as to induce a belief that scientists
must be born, or at least made, as superior beings. This picture of
the perfect morality of science had its strong basis in reality, both in
 Social Problems ofIndustrialized Science 63
the practice and in the self-consciousness of science in the academic
period. For then, the favourable social conditions, and the effective
arrangements for the protection of property, were sufficient to
ensure that the working ethics of science would serve to resolve all
the social problems encountered in its experience.
In the longer perspective of history, the moral innocence of
academic science appears as a temporary feature, a happy accident
of circumstances. For the problem of responsibility for powerful
knowledge had been recognized by the practitioners of fields ances-
tral to science, in all the previous centuries. Such knowledge,
including magic, alchemy, and parts ofastrology, was to be restricted
to those who could use it wisely; and hence it was transmitted in an
oral or cryptic tradition to initiates in a brotherhood. Of the pioneers
of the scientific revolution, Bacon and Descartes, although dis-
believing in magical powers, retained this moral sense. Even in his
Utopian 'New Atlantis', Bacon had the sages of 'Solomon's House'
deciding which secrets they would reveal to the State, and which not;
and Descartes stated a 'scientist's oath' of classic simplicity: I would
not engage on projects which can be useful to some only by being
harmful to others. 38 But Galileo, whose general style of work was so
much more like that of a 'scientist' than any other natural philoso-
pher of his time, was totally lacking in such a sensitivity. He con-
sidered himself as having the right to proclaim philosophical truth
as he saw it, and was utterly unconcerned with the possible social
effects of his unsettling doctrines. It is probable that he was sure
that God's truths, which he was announcing, could not be harmful;
but in practice he was demanding the influence over men's minds
resulting from his pronouncements, while denying responsibility for
the consequences of his actions. His moral position was made even
more complex by the fact that while fighting hetoically for the truth
38 For Bacon: 'And this we do also: we have coDSUltations, which of the inventiODS and
experiences which we have discovered shall be published, and which not: and an take
an oath of secrecy, for the concealing of those which we think fit to keep secret: though
some of those we do reveal to the state, and some not.' NI1IJ Atliln,;s, in Bacon's W"i"
eds. Spedding, Ellis, and Heath, 14 vols. (London, 1857-74), VoL 3, p. 165. UnJessother-
wise noted, aU references to Bacon will be taken from this edition. The American version
of this edition unfortunately has a different pagination.
For Descartes: ' ••• et que mon inclination m'eloigne si fort di toute sort d'autra
desseins, principalement de ceux qui ne sauraient atre utiles au uns qu'en nuisant au
autres, Ii quelques occasions me contraignaient de m'yemployer, je ne crois point que je
fusse capable d'y reussir.' Distows de 1tI MetlUJde, last page of text, in the edition
Distows de I. Methode, texte et commentaire par E. Gilson (Vrin, Paris, 1925, 1926)
p. 78. Unless otherwise noted, references to the DistowS will be taken from this edition.
64 The Varieties ofScientific Experience
of natural philosophy as he saw it, he was not at all averse to making
money privately through the sale of his inventions to the highest
bidder. 30
With the rise to dominance of the 'mechanical philosophy', or
more correctly with the dehumanization and disenchantment of
Nature for the educated common sense of European civilization, the
expected powers of scientific knowledge were reduced, and hence
also the responsibilities of men of science. The association of natural
science with 'progress', first with that ofthe intellect in the eighteenth
century, and then with industry in the nineteenth, removed any
fears concerning the applications of science from the traditions of
self-awareness within science. Hence it was as moral innocents that
a group of distinguished physicists urged the American government
to produce an atomic bomb. Although such a device clearly had its
dangers, ordinary human morality dictated that there should be no
risk of Hider having sole possession of such a bomb. But, some six
years later, what had been conceived as a deterrent weapon in the
struggle against Fascism, was used on the civilian population of a
nation near to surrender; and not once, but twice. Were the scien-
tists who initiated the Manhattan Project morally responsible for
the victims of Hiroshima and of Nagasaki ?40 This is not an easy
39 Of the many accounts of GaliIeo's life and struggles, the one which best indicates
the mom problems of this sort is J. Brodrick, RolJert Belltmnine, StIi", lind S,lJo14,.
(Bums & Oates, London, 1961), ch. 12.
GaliIeo twice tried to sc11 his method of determining longitudes at sea by observation
of the eclipses of the moons ofJupiter. The first attempt started in 1616 with Spain, and
GaliIeo even offered to travel there to train practitioners. The second attempt was with
the Dutch States-General in 1636, a few years after protracted negotiations with Spain
bad collapsed. See J. J. Fahie, GfIli/eo, llis Lift lind W.i Uohn Murray, London, 1903;
reprinted Wm. C. Brown, Dubuque, Iowa, 196'7), pp. 173-7 and 372-5. In each case, the
prize offered by the foreign State for a succesaful method overcame any scruples that
GaliIeo might have bad in dealing with them: the one arch-reactionary Catholic, the
other heretic; and an this during the Thirty Years' War. I am indebted to Mr. Peter
Buck, then a research student at Harvard University, for the discussions in which this
interpretation of GaliIeo'. moral problems became clear.
40 To my knowledge, the full history of the decision to bomb Nagasaki u well as
Hiroshima has never been told. One theory is offered by Gar Alperowitz: ' ..• I think
the only way you can understand why Nagasaki was tripped off, automatically, bing-
bing, just like that, with no consideration, is this tremendous rush to end the WH-
. , just to end the war before an invasion, but m..ttlitlte1l/ ••• What was the rush?
Well, P. M. S. Blackett, another Nobel prize winner, saw in 1945 that the only way you
could explain that immediate, fast one-two punch, was the fact that the Russians were
in fact scheduled to enter the war on August g. And it's in that context, to end the
war, not just before an invasion, but bam, like that, that you explain Nagasaki on
August g.' See J. Allen (ed.), M""IJ 4, S,imtisls, SlrulmtS, 11,,4 Sonn, (M.I.T. Press,
1970 ), p. 174-
 Social Problems ofIndustr;alized Science 6S
question to answer; but whatever the judgement, it is clearly not one
which can be reached by an argument within the traditional working
ethics of science. 4 1
Thus, first with the Bomb, and more recently through involvement
in runaway technology, the world of science has been faced with
genuinely moral problems. They concern the responsibility of an
individual for the immediate and remote effects of his actions. Such
problenlS are difficult, if not insoluble. The inherited working ethics
of science offer no guidance; and the traditional claim of benefit for
humanity has, in its realization, produced this darker side. Whatever
else will happen to the ideology of science, it can never again claim
Innocence.
A few decades is a moment in the life of a civilization, and does
not even cover the working life of a scientist. The sudden transfor-
mation ofthe social activity ofscience, rendering its traditional ethics
obsolete and its ideology hollow, has caught many a scientist in a
personal tragedy. That ofEinstein is the most famous; the incidental
relation E = me", arising from his profound studies in the philosophy
of nature, became the magic fonnula for the sorcery of the atomic
bomb, for whose creation he was at least partly responsible. There
are doubtless many other such cases: the President ofthe University
of Pennsylvania, formerly a distinguished physicist, found himself
in an unpleasant situation when, after repeated denials, he was forced
to admit that germ-warfare research was being conducted secredy on
his campus. Calling for his resignation, the student newspaper
explained that, as an elderly man, he found himself involved in
problems whose existence he could not have conceived when, as an
old-fashioned scientist, he took office some years previously. 42
41 According to his account, J. Bronowski conceived his book Seimee .u HIItM"
Y.lws (Hutchinson, London, 1961) in response to the grotesque experience ofthe ruins
of Nagasaki seen to the accompaniment of American popular music coming from a ship
tied up at the harbour. He achieved an eloquent, and indeed classic, statement of the
creativity and nobility of the best scientific en«:cavour. Unfortunately, he was unable to
solve the moral problem in this fnmework; concerning the evil effects ofatomic physics,
he could say only, 'Science has nothing to be uhamed ofeven in the ruins of Nasagaki.
The shame is theirs who appeal to other values than the human imaginative values which
science has evolved' (p. 83). On the question of why these values gave so little guidance
to those who stayed with the Manhattan project through to its culmination after the
defeat of the Nazis, and who later proceeded to the construction of hydrogen bombs, he
is silent. In fairness, he did try to convince oflicial agencies to have Nagasaki preserved in
its pristine ruinous state, IS a site for international conferences on peace; but the idea
was clearly too sensible to be adopted.
42 See editorial, 'Hamwell-an Old Man', TAe Dili/] Pe""s,/w";." (~8 April 196'1),
66 The Varieties ofScimtijie Ezperim&e
For many generations, up to but not including the present, the
study of nature has been among the most serene of occupations. To
be sure, the work is arduous and even hazardous; but many an
eminent scientist turned with relief from the turbulence and faction
ofpolitics, commerce, and even institutional religion, to the innocent
contemplation of the unchanging and impersonal laws of nature. 43
In less than a generation's time, that haven has been lost. Science is
in flux. Many who entered it as a refuge from the intellectual and
moral squalor ofordinary society find, in their advancing years, that
they are involved in administering just another bureaucratic estab-
lishment. They are enmeshed in the demands of society and the
State; they must accomplish the administrative and social tasks of
getting high-quality craftsmen's work out ofa set of manpower-unit
employees; and in participating in the leadership of their field, they
must cope with the insoluble practical and moral problems which
emerge when corruption sets in.
For the industrialization of science has produced another set of
moral problems, internal to science but still incapable of solution
through a working ethic that was conceived in terms of the search
for truth, and organized around the protection of well-defined
intellectual property. As the ideal of truth has become obsolete, and
the location of intellectual property has shifted, this ethic hac; lost its
relevance. The very naturalness of such conditions as shoddy
science and entrepreneurial science ensures that they are connected
to good work by a multiplicity ofcontinuons gradations. A scientific
entrepreneur may produce a piece of genuinely good work, but then
p. .. I lID indebted to Mn. E. Brown of PbiJadelpma for her help in locating this
reference.
43 The tatimoDy ofEiDstein in this regard bas a particular pathos: 'It is quite clear to
me that the religious paradise of youth which was thus lost was a first attempt to free
myself &om the chains ofthe "merely-personal", &om an existence which is dominated
by wishes, hopes and primitive feelinp. Out yonder there was this huge world, which
exists independently of us human beiDp and which IIaDds before us like a great eternal
riddle, at least partially accessible to our inspection and thinking. The contemplation of
this world beckoned lib • liberation, and 1 SHII utieetl,Ut """1 II ",." ",/wm I IIIIIl
Iellt1letl to esteem IItIIlllllmire IuIIl fOlWl;""" fteetlfntt IIfIIl senlrit1;n tM tlevotetl OC,"p.t;on
",;,11 it. The mental pup of this ema-pencmal world within the fnme of the given
p0S81oilities swam I I highest aim half consciously and half unconsciously before my
mind'. eye. Similarly motivated men of the present and ofthe past, u well as the insights
which they bad achieved, were the friends which could not be lost. The road to this
paradise was not I I comfortable and alluring I I the road to the religious paradise; but it
has proved itIe1CI I trustworthy, and I have never regretted having chosen it.' A. Einstein,
'Autobiographical Notes' in AlNn Bigli", PllilDsplwr-S&inllist, ed. P. Schilpp (Tudor,
New York, 1 949, 195 1), p. 5 (italics added).
 Social Problems ofIndustrialized Science 67
use it for further inflating the stock of his establishment; has he
violated the ethics ofscience? A piece ofwork may be condemned as
shoddy, but may then be defended as the best work a particular man
could do in a difficult field; the assessment of quality of scientific
results is subde enough, and who can be sure of the intentions of
another? And if a man does not wish to belong to a community of
colleagues of the traditional sort, and does not need it for the
building of his personal career in science, how can the informal
penalties of disapproval be applied against him? Thus there is no
escape from moral problems in science, even in the purest of pure
science, except perhaps in those enclaves which have not yet been
affected by industrialization and its social consequences.
Conclus;on
It might seem far-fetched and alarmist to claim that science is in
danger of decline and dissolution, through its inability to make a
healthy response to its new conditions. The size of the whole enter-
prise, and the ever increasing number of worthwhile and exciting
results which appear, seem to be a patent refutation of any such
fears. But the history of natural science in Europe shows that its
steady growth over the past centuries has been an aggregate ofcycles
of growth and decline in different fields and places. 44 Indeed, it
would be astonishing if it were not so, for then natural science would
be the only sort of creative work exempt from such rhythm. 45 Only
44 See M. Yuasa, 'Center of Scientific Activity: its Shifts from the 16th to the 20th
Century', ,_paMse Studies in 1M Bistor, ofSdmee, I (1g62), 57-'75. The author used
several sorts of quantitative indices of 'activity', including the mean age of scientists at
any time. They agreed in producing cycles ofabout eighty years in length for each nation;
that for France was at its peak at the time ofthe French Revolution, that for Germany in
the 18705; that for America seemed near to its peak after the Second World War, and
that for Russia was still rising. Detailed historical studies of France and Germany
confirm this analysis. For France, see R. Fox, 'The Rejection of Laplacian Physics: a
Turning-Point in the History of the Physical Sciences in France', to be published in
hellive for Histor, oft lie Ex"' Sdmees; and for Germany, see J. Ben-David and
A. Zloczower, 'Universities and Academic Systems in Modem Societies', Eurole""
J01InIIIl of Sodology, 3 (1g62), 45-85, reprinted in N. Kaplan, Slime, "fill SOlin"
pp. 6~-85·
45 A striking example of the appearance of a creative generation is given by G. S.
Shackle for economics: 'The twelve years from 1840 to 1851 produced Menger (1840-
1921), Marshall (1842-1924), Edgeworth (1845-1926), Pareto (1848-1923), Wicksell
(1851-1926), Wiesser (1851-1926) and Bohm-Bawerk (1851-1914), seven of the greatest
figures of our discipline, all bom in virtually one decade, all but one dying in the six
years 1921-6. By that last year, which we have taken as the first of our Years of High
Theory, the great Victorian cohort had at last withdrawn into antiquity. A fresh start
could be made without these giants peering over men's shoulders. Thus need and ftee-
dom beckoned.' See Tht Years ofBig" ThtDr] (Cambridge University Press, 19&]), p. 296.
68 The Varieties ofScientific Experience
if one shares the nineteenth-century faith that the advancement of
knowledge is automatically progressive and cumulative, can one
believe scientific inquiry to possess the unique combination of
features of being highly creative and also perfectly safe. But once the
existence of such cycles is recognized as a natural occurrence, the
close succession of periods ofexcellence in different fields and places
becomes an entirely contingent process: there is no guarantee that in
the later twentieth and twenty-first centuries there will always be
some national centre of excellence in science. 46 Only in retrospect
can historians begin to explore the subde combination of factors
which led to the inception and completion of a cycle of creativity in
any given situation in the past; for the present and the near future,
when conditions are so very different from those experienced
previously, we really do not know.
Thus in spite of its great achievements, and indeed because of
some of them, the world of natural science faces serious problems of
an entirely new order. For these social problems, neither the
inherited craft wisdom of working scientists, nor the fledgling social
sciences, are in a position to provide solutions. Indeed, the social
sciences themselves, and sophisticated technology as well, have
analogous problems. The solutions, if there be any, cannot be
imposed from outside by the fiat of politicians or administrators;
scientific inquiry is too complex and delicate to be treated like the
production of material commodities. What is required is first under-
standing, and then, above all else, the intellectual and moral qualities
of a new leadership, capable of adapting the best of the heritage of
science to the tasks of the present and future. Whether the present
conditions are propitious for the emergence ofsuch a leadership will
be known only by those who look back from the future.
46 A very emphatic graphic display ofan earlier gap in scientific creativity is given on
the back inside cover of A. E. E. McKenzie, Tile M.jrw Acllitwmnlls of Snmce, vol 1
(Cambridge University Press, 1960). He has a bar-clwt of life-spans of great scientists;
there is a cluster ofeight between Pascal (born 1623) and Newton (born 1642); then after
Stahl (born 1660) there is nothing until BWfon (1707). Although such selections of
names are bound to be arbitrary, the impression created is a correct one j and the decline
in recruitment and enthusiasm for experimental philosophy was recognized at the end of
the seventeenth century.
 Part II
THE ACHIEVEMENT OF
SCIENTIFIC KNOWLEDGE
 INTRODUCTION

AT the present time, the deepest problems in the understanding of

science are social rather than epistemological. The older problems
of the attainment of truth or its substitute, have given way to the
concern for the maintenance of the health of science, and for the
control of its applications. Until very recendy, the two aspects of
science, the activity itself and the resulting knowledge, have been
studied separately. Both sorts of inquiry have been impoverished by
this separation. Analyses of the social behaviour of scientists, and of
the external influences on scientific research, have assumed the
products of that research to be absolute, and unconditioned by the
particular circumstances of their achievement. On the other hand
philosophical analyses of the nature of scientific knowledge have
either been completely abstracted from the processes of its achieve-
ment, or have invoked a model of a working scientist isolated from
his environment and traditions. But a proper analysis of the social
activity of science must be based on understanding of the very
special goals of the scientists' tasks; and an analysis of achieved
scientific knowledge must comprehend its character as a social
possession, the product of an historical process. An analysis of
science which unites these two aspects will be able to resolve the
apparent paradoxes in its nature: that out of a personal endeavour
which is fallible, subjective, and strictly limited by its context, there
emerges knowledge which is certain, objective, and universal. Also,
such an analysis should help to explain the failures of scientific in-
quiry as well as its successes, and could provide guidance for the
solution of the social problems of the science of the present and near
future.
This part of the book will be devoted to the problem of scientific
knowledge. Its argument will be organized around four theses:
I. Scientific inquiry is a craft.
2. The objects of this work are not natural things, but are in-
tellectual constructs, studied through the investigation of
problems.
72 The A,IJievemmt ofScientijie KnoJ1)ledge
3. The work is guided and controlled by methods which are
mainly informal and tacit, rather than public and explicit.
4- The special character of achieved scientific knowledge is ex-
plained by the complex social processes of selection and trans-
formation of the results of research.
On the basis of these four theses, it will be possible, in the next
part, to discuss some problems of science which arise from its being
such a delicate and w1nerable social activity.1
In the discussion of these four theses, the central concept will
be that of 'problem'; I shall consider science as a special sort of
problem-solving activity. This concept cannot be left as a common-
sense idea needing no close analysis; the work it does in this argu-
ment is too important. As my argument proceeds, I will develop a
general schema of the distinct phases in the setting and solving of
scientific problems. Indeed, this part could be organized around a
formal doctrine of 'scientific problem', but the gain in coherence
would be more than offset by the difficulties presented by a battery
of unfamiliar terms. For this reason I shall defer the formal defini-
tion of 'problem' until it can be meaningfully discussed. For the
present, we may think of a scientific problem as analogous to a text-
book 'exercise', with the following crucial differences: a major part
of the work is the formulation of the question itself; the question
changes as the work progresses; there is no simple rule for dis-
tinguishing a 'correct' answer from 'incorrect' ones; and there is no
guarantee that the question, as originally set or later developed, can
be answered at all.
There is one important similarity between a scientific problem
and a textbook exercise: the things discussed are not the objects and
processes perceived through ordinary experience, but intellectual
constructs. This is easily seen in the case of the more mathematical
and theoretical sciences; the schoolboy solves problems about
'mass-points' which are considered to have no size, while the
theoretical physicist discusses things whose properties can be known
only by the most indirect and sophisticated experiments, and whose
very existence is sometimes a matter ofcontroversy. I shall argue that
I A condensed and simplified version of the ideas developed here will be found in the
latter put of Course Unit I of the Foundation Course in Science of the Open University.
I lID paIeful to the regular members of the Course Team for their aiticisms and im-
provementl of the dnft there, and only regret that I could not enlist their help in the
impI'ovemeot of this text.
 Introduction
73
all disciplined inquiry is necessarily concerned with objects of this
sort; even in the most 'descriptive' fields, such as ordinary history,
the study of particular events is organized in terms of concepts as
'nation', 'class' and 'progress'. I describe these 'objects of inquiry' as
'classes of intellectually constructed things and events', rather than
as 'concepts', in order to stress what, for me, is their most important
feature. It might be objected that science sometimes discusses
unique things apprehended by ordinary experience, such as the
moon. But if one reads scientific discussion of the moon, one sees
that all the properties discussed are just the sort of intellectual con-
structs that physics deals with. Indeed, the 'common sense' per-
ception of the moon was transformed by a brilliant exercise in
scientific research, using inspired observation and careful, if in-
formal demonstration. Before the appearance of Galileo's Sidereus
Nuncius of 1610, everyone knew that the moon is a perfect sphere,
with some unexplained markings. With his telescope, Galileo ob-
served that the moving boundary between the light and dark parts
is not regular, but broken; and there were changing patterns of light
and dark spots on the dark and light parts respectively. He inter-
preted these creatively (and dangerously) as shadows, and produced
a geometrical argument to show that they are cast by mountains not
impossibly high.
My discussion of the successive theses will run roughly parallel
to the description of the different phases of the investigation of
scientific problems. For the craft character of the work is seen most
clearly in the earlier phases, where there is an interaction with the
external world; while the artificiality of the objects of inquiry is most
obvious when we consider what is involved in 'solving' a scientific
problem; the social character of scientific inquiry is exhibited in the
methods that guide and control scientific work; while its dependence
on social processes operating over time, is shown in the transforma-
tion of 'research reports' of solved problems into accepted 'facts'
and ultimately to genuine scientific knowledge. 2
2 In the philosophicallitenture conceiving science IS inquiry nther than IS valida-
tion of accomplished knowledge, a pioneering effort which deserves to be remembered
is F. C. S. Schiller, 'Scientific Discovery and Logical Proof', in C. Singer, SlIUlies in tile
History (,,,,,1 Method ofSdmee (Oxford, 1917). For the recent studies in this direction,
the work of T. Kuhn, The Stn«ture of Sdentific Revolutions (University of Chicago
Press, 1962) has, IS it were, created the new pandigm which we an foUow. Complo-
mentary to Kuhn is J. Ziman, Public KtIOJI)letl,e (Cambridge University Press, 1968),
in which the social ISpecb of scientific work are given greater emphasis. The concept of
74 The A,hievement ofS,;entiji& Knowledge
In the early parts of this discussion, I will show in great detail
how the pursuit of scientific knowledge is, in the short run, very
subjective and fallible. It may even appear that I am arguing towards
a completely sceptical position, with the practical conclusion that
scientists should only attempt to 'solve problems' and not concern
themselves with the goal of achieving knowledge. This is quite
contrary to my intention; I start with the historical fact that
genuine scientific knowledge can be achieved, and does exist. The
problem is then to see how this occurs, given the common experience
of the uncertainty of scientific research. It is necessary for me to
elaborate on the sources of uncertainty, both for the correction of
the bias implicit in the philosophy of science dominant hitherto,
and for the provision of materials for the solution of the main prob-
lem. This solution emerges at the end of the section; it is that the
complex social processes of the testing and transformation of the
results of research, working over an extended time, create both
genuine scientific knowledge and also the circumstances in which it
can be recognized. The paradoxical properties of this knowledge
(that it appears in a variety ofdifferent versions, and contains hidden
obscurities) can then be appreciated and explained.
'problem' is studied at length in]. s. Sbarikow, 'Du wisscuschaftIic Problem', in
IA,ij W JI1Wnu&1Mjili&1Im FfWl&"', eels. P. W. Kopnin and M. W. PopowitlCh
(Abdanie-Verlag, Berlin, 1969; !lUISian on,iDal, Mosmw, 1965), cit. I. His analysis
and miDe are remarbbly similar OIl fundamentals, although the materials of my cbs. 3,
4 and 5 diverge &om his approach. Mario Bunge, $dnlliji& Rese.eIJ 1: Tie Se.eIJffW
Sy"mt (SpriDpr-VerIag, 196'7), bas a discussion of 'problem' in ch. 4 and offen many
Ipecial iDsighta.
 3
SCIENCE AS CRAFTSMAN'S WORK

To anyone with experience of the 'art' of scientific inquiry, this

thesis may seem so obvious as to be banaI. 1 Yet this feature of
science has generally been ignored in philosophical discussion, even
in those that try to take into account the work by which scientific
knowledge is achieved. 2 Yet without an appreciation of the craft
character of scientific work there is no possibility of resolving the
paradox of the radical difference between the subjective, intensely
personal activity of creative science, and the objective, impersonal
knowledge which results from it. When we think of material objects
produced by handicraft rather than by mass-production, we easily
appreciate the distinctive features of this sort of work. The crafts-
man works with particular objects; he must know their properties in
all their particularity; and his knowledge of them cannot be speci-
fied in a formal account. Indeed, no explicit description of a crafts-
man's techniques, and ofthe objects on which he works, can be more
than the simplest elements of the subject. They can be useful for
the beginner, but he must develop a personal, tacit knowledge of his
objects and what he can do with them, ifhe is to produce good work.
Indeed, much of his technique may not even have the character of
IOn the craftsmaDship or science, see w. I. B. Beveridge, TIle Art of SMllifi&
IftWsd,tIIiIm (Heincmum, ~ 1950), and Paul Freedman, TIw Pritleilln of
Sdnlliji& ReseII',IJ (perpmon Press, 1960). An account which provides a walth of
practical aample and aphorisms is M. HamiltoD, LeawaM ,lit MnluNlDltv o/CIirW.J
Resell"" (E." S. Livinptone, Edinburgh and London, 1961).
a I am indebted to Mic:bae1 Pouyi for his I)'ItaDatic development of the iDsight
that science is craft: work; see PerSOlUll K.1I1W,e (Routledp " Kepn Paul, Loudon,
1958). Much or this present work derives &om an attempt to solve the problems which
were nised by his analysis; in particular, how objective scientific knowledge em result
&om the intensely personal and fallible endeavour or creative scientific inquiry. AIIo,
reflection on D. S. L. Cardwell's distinction between 'craft' and 'applied science'
(exemplified in piIocage and navigation, respecdftly) helped me aystaIIDe my own
idea on the craft aspects of scientific work.
76 The Achievement ofScientific Knowledge
conscious knowledge; by experience, his hands and eyes have
taught themselves. It is this subtle interaction of the craftsman with
his material, producing slightly different copies of the same general
model, which gives handicraft productions their special charm.
At first sight, nothing could seem further from such productions
than achieved scientific knowledge. Those results which reach a
wide audience bear few marks of the individual hand and mind
which brought them into being. Indeed, those which have become
'classic', have an appearance of such simple necessity, that it is
frequendy hard to imagine that their achievement required the
exertions of a great talent. Working from such materials, it is very
difficult for the historian to argue convincingly that their production
was conditioned by such contingent factors as the social or intellec-
tual environment of their creator, or by peculiarities of his per-
sonality. This apparent paradox will be fully resolved only through
an analysis of the transformation of scientific results as they pro-
gress towards the state of being permanent knowledge. But the
interplay of personal and social aspects of the work begins with the
birth of the problem itself and is continuous through the inquiry. I
will defer the discussion of the origins of problems for the next
chapter; for the present I will show how the materials from which
the solution is eventually constructed are produced and transformed
by operations which require a craftsman's knowledge of the work.
D4tll
As a first example, let us consider an experiment in which quanti-
tative readings are taken from a piece of experimental apparatus.
Now, one of the things that every schoolboy knows about science
is a general property of experimental equipment, which has been
given the name of the 'fourth law of thermodynamics': no experi-
ment goes properly the first time. The schoolboy knows that the
equipment is not functioning properly, from its failure to produce
readings of the required sort. If by some miracle it performs to
expectations on the very first run, he can happily record the results
in his laboratory notebook, and go home. For him, the situation is
quite straightforward: everyone knows what result should be
achieved; until it comes, the apparatus is not working properly; and
when it does, the apparatus is working properly.
This simple and secure rule of assessment is not available when
new work is being done on a piece of experimental apparatus. How,
 Science as Craftsman's Work 77
indeed, does a scientist decide that his apparatus is working
'properly'? Of course, hardly any experiment is completely new,
and so the scientist will always have an idea of what to expect. If
nothing else, he can expect a set of readings which will produce
some sort of regular pattern; and if they gyrate wildly, then he
does not need much craftsman's knowledge to realize that the
equipment is not yet functioning properly. But when to say that it
is functioning 'well enough' is a more subtle matter. It is clear that
the more stable and consistent the readings, the more likely they are
to be sound. But anomalous readings always do crop up; and if one
waited for them to vanish entirely, or tried to 'explain' each and
every one of them, one would never get beyond this first stage of the
work. In short, the scientist must be a craftsman with respect to his
apparatus; and his judgement of when it is working 'well enough'
must be based on his, experience of that particular piece of equip-
ment, in all its particularity.3
In this work with pieces of physical equipment, the scientist is a
very special sort of craftsman, for the objects he is dealing with are
highly artificial. The relation of the readings taken off the apparatus
to the objects of his inquiry is not at all immediate: the establish-
ment of their relevance requires another set of operations. The
experimental apparatus itself is frequendy a complex affair, de-
signed for the production of very special effects. A purely 'practical'
mastery of the technique of standard manipulations will be suffi-
cient only for the most routine work. Without a deeper knowledge
of the operation of the apparatus, the scientist may well fall into
the first pitfall of experimental research, that of too easily accepting
readings which are stable for reports which are sound. For the
readings taken off the machine are only a measure of the response
of the output mechanism to certain aspects of the process going
on inside. At every stage of what may be a lengthy sequence, the
apparatus itself will modify the signals coming through, and so in-
troduce a 'systematic error' into the readings. The scientist must
know by experience what size of error is 'negligible'; he must be
able to reduce the recognized errors to such a size; and he must
be sensitive to clues announcing the presence of stable
3 There is a lengthy discussion of 'Bringing an Apparatus under Connol' in E.
Bright Wilson, Jr., An Introtluetion to S&ientiji& Rese.eIJ (McGraw-Hill, 1952), pp.
137--40; but there is no mention of the aiteria for judging that the apparatus ;s under
controL
78 The A&lUevnnmt ofScientific Knowledge
non-negligible errors from other sources. For this sort of work, the
scientist must be a master of the operations of the apparatus. Such a
mastery comes partly from craft knowledge of the traditional sort,
where the experience of others is transmittetl by precept and imita-
tion; but it also involves some explicit and formal scientific know-
ledge in which the 'theory' of the apparatus is set out. Because
of this formal component of his knowledge of his apparatus, the
scientist must be a 'technician' in this respect rather than a 'craftsman'
of the traditional sort.
It is clear that the set of quantitative readings taken off the experi-
mental apparatus cannot be considered independendy of the
interpretation put on them; the scientist cannot even deal with the
systematic errors of the apparatus unless he has some idea of what
the readings represent, as reports of properties of particular things
and events inside the apparatus. It is such reports, rather than the
bare set of numbers, that are the objects of this part of the work;
we may call them the 'data'. We shall soon see that there must be a
further series of operations on these data to yield materials suitable
for inclusion in the argument of the solved problem, and the pro-
duction of the data must be accomplished with these subsequent
operations in mind. But the data are a record of the point of contact
with the external world, and as such constitute the foundation in
experience of scientific knowledge.
A different set of craft skills are involved in the production of
data in fields which are 'descriptive' rather than 'experimental'. In
these, we may say that the data are 'found' rather than 'manufac-
tured', for the things and events whose properties are reported are
not studied in artificially pure, simple, and stable situations, but
are instead a sample taken from an existing population. Data in such
fields cannot enjoy the precision of that which comes from experi-
ments, for the interaction with the external world cannot be so
closely controlled. Accordingly, such data cannot serve as the
foundation for evidence in a highly abstract and subtle argument,
and so the properties which are reported will tend to be more similar
to those accessible to common observation. This does not obviate
the need for craft skill in making reports, for there are pitfalls in
these data as in the other sort. For the process ofsamplingis governed
by the circumstances of the scientist and the tools at his disposal,
and can quite easily yield specimens which are unrepresentative of
their population in some important aspect.4 Also, any scientific
 Science as Craftsman's Work 79
description ofa property is necessarily abbreviated and stylized; and
the existing categories of description will never fit perfectly the ob-
jects of the inquiry at hand. s Hence here, as in experimental fields
it is necessary for the scientist to have a craftsman's mastery of his
materials and tools, and to be able to judge when a report is 'suffi-
ciently sound' for the data-collecting to be concluded, and the re-
finement of the data into information to commence.
It might appear at first that in theoretical or purely mathematical
fields, where there is no interaction with the external world at this
early stage, the category of 'data' could not apply. But mathe-
maticians will also speak of 'experiments' and 'inductions' in the
early stages of working up a proof. 6 In this case the objects are
4 In fields where the data are found, the problem of coping with 'oudying' data is
particularly severe, for one can only nrely repeat the process under observation. One
might tty to use statistical tests to estimate the significance ofsuch data, deciding whether
they are indications of a regular cause, or merely freaks; but all such tests depend on
hypotheses about the universe from which the data are drawn, and so are less likely to
be genuinely applicable in this extreme case. The extent to which such oudying data
occur in experimental and observational work is likely to be underestimated by those
unfamiliar with such work, since it is frequendy suppressed. A most striking example of
such suppression (in the teaching literature) is the classic study on deaths by horse-kicb
in German army corps, which is the paradigm case for the Poisson distribution, describ-
ing randomly occurring events. The original study of horse-kick deaths by von Bartko-
witseh extended over fourteen army corps, but the data from four of these was rejected
as being anomalously high, and the data from the remaining ten was shown to fit very
well the theoretical numbers. A note of this suppression remains in some of the litera-
ture; see ]. S. Coleman, Introduet;tm to M.thmulti&.l Sotioloo (Collier-MacMillan:
The Free Press, Glencoe and London, 1964), p. 291. But this is the exception; there is
no hint that the horse-kick deaths data is other than 'raw' in R. A. Fisher, St.ti,,".'
Methodsf. Rest.,1J W.iers, 13th ed. (Oliver & Boyd, Edinburgh and London, 1955),
P·55·
5 For an example ofthe subdeties of 'descriptive' sciences, one may consider the sorts
of problems encountered in constructing a bibliography. The first decision is on the
'field'; and even here the existing conventions must always be modified by the par-
ticular circumstances or even because of the passage of time. The taxonomy of classes
is an obvious source of trouble. Then, should it be 'comprehensive' or 'critical'? In
pure versions of the former, it will be choked with obsolete and inferior material and
is likely to be an unmanageably large project; in the latter it will be 'subjective', in-
complete and possibly misleading unless the principles are clear. When worb are cited,
should .U editions up to the present be mentioned; and, if not, what principle of se1ec-
tion should be used? Even on the apparendy ttivial question of the alphabet, there are
decisions to be made on such special cases as 'Me' and the German umlauts. There is
one large university library where the German 'ii' is treated as 'ue'; and the French
name 'Hauy' is filed as if it were 'Hauey' lin pnctice, unless there is some prior decision
on the f~tm of the information in a bibliography, these problems are resolved (if
at all) only as the bibliographer becomes aware of them singly, and the result is likely
to be a mass of data which gives satisfaction in no aspects whatever.
6 See G. Polya, M.,1JmuJIjes tUUl PJ.1UilJ1e RelUMJi"l, voL i (Oxford University Press
and Princeton University Press, 1954): l"dll&t;01I .tul AuloO;" M.thmulti&,.
80 The Achievement ofScientific K"oT1Jledge
purely mental constructs; but they will be more simple or more
familiar representatives of the general classes with which a mathe-
matical proof is usually concerned. An exploration of the properties
of these samples, in relation to the partly-formed conjecture with
which the mathematical investigation begins, gives the mathemati-
cian some idea of the content of the result he should try to establish,
and also hints towards the sttucture of its proof. Indeed, in many
cases the proof emerges from a repeatedly refined and generalized
description of the properties of the sample objects.
A craftsman's intuitive knowledge of these absttact objects,
derived from experience in manipulating with them in a variety of
contexts, is essential for such 'data' to be a sound starting-point for
the detailed investigation. Without a sense of the generality implicit
in his sample objects, and without a sense of the possibility of
refining his descriptions into rigorously argued components of a
complete proof, the intending mathematician will wallow in a
morass of ill-formed speculations and partly solved conjectures,
unable to develop towards a tight, coherent, and general argument.
Here too there are pitfalls in the work, which can be avoided only
by the craftsman's intuitive knowledge of his objects. For the work
of constructing a mathematical argument cannot proceed other than
by the testing and development of a rough conjecture, based on
this special sort of experience. At a certain point the mathematician
must decide when this 'data' is sufficiently sound to justify the in-
vestment of time and intensive effort for its refinement. There can
be no 'proof' that the sample objects are ttuly representative, and
that the descriptions of their properties reflect the structure of the
general classes. Moreover, every sketch argument involves the use of
ancillary and special results, which seem either trivially ttue or
easy to prove. Hence it can easily occur that plausible objects, with
plausible properties, and with obvious 'lemmas' supporting the
argument, tum out to be unrelated to the objects with which the
intended proof is concerned, or indeed to any significant class of
mathematical objects whatever; but this can be discovered only by
disheartening experience, when the process of consttucting a proof
fails to 'converge', and the mathematician is left with a debris ofinsig-
nificantspecial exampl~ scattered among broken-backed arguments. 7
, Eumples or this phenomenon are not easy to find in the worthwhile published
literature; but I em cite my own apaieDce .. a beginning raearch student, and that
offiienda.
 Science as Craftsman's Work 81
Thus in any sort of scientific inquiry we may use the idea of'data'
for the results of the first working-up of the materials in the in-
vestigation of a problem. I have stressed the importance of craft
skills in the production of data which is sound of itself, and useful
for the later stages of the work on the problem. Without such skills,
the scientist will blunder into pitfalls and produce reports without
soundness or significance. The craft character of the production of
data is of some philosophical significance, for it is in this phase that
the scientist makes new contact with the external world, or achieves
a new organization of his conceptual objects. The directions and
fortunes of the later stages of the work will depend on the results of
this work. Yet we have seen that no set of data can be 'perfect' as a
report of properties of the objects of investigation, nor can it be
independent of the plans and expectations for the later stages of the
work. If such reports' are ttuly the foundations of scientific know-
ledge in experience of the external world, then those foundations
would seem to be peculiarly insecure and complicated. And so it is;
the wonder is not that our scientific knowledge is an imperfect and
fallible picture of the external world, but rather that it exists at all.
In the light of the history of human inquiry into the natural world,
we see how difficult it is for such knowledge to come to be. I shall
later show that when such knowledge does exist, it is achieved by a
complex social endeavour, and derives from the work of many
craftsmen in their very special interaction with the world of nature.
We have already had an intimation of the social character of the
activity of science, in the judgement made by the scientist in de-
ciding whether a batch of data is 'sound'. For the standard he
applies will derive partly from his own experience, and partly from
that of his teachers and colleagues. Since perfection of data is im-
possible, the standard will be based on a common judgement of
what is good enough for the functions which the data performs in
problems ofthat sort. This judgement depends in turn on the criteria
of adequacy imposed on the solution of such problems. The solu-
tion cannot yield necessary ttuths or indisputable facts, but only
results which in their tum are judged acceptable (or not) by the
community in accordance with criteria developed through its social
experience. To be sure, some ofthese results may, either themselves
or their descendants, come to be classed as 'facts' or even 'know-
ledge'. But Nature is not so obliging as ever to give marks of True
or False for scientific work, and so in the last resort a scientific
82 The A&/Uevemmt of s&imtiji& K"oT1Jledge
community must set its standards for itself: The sets ofsuch standards
over the different fields are very diverse in their contents and rigour.
The self-discipline of scientists which they require, and the quality
of the work they ensure, depend on the ~ength, health and in-
tegrity of the community involved, and are most sensitive to the
leadership it receives. 8 Thus the simple judgement of 'soundness'
of data is a microcosm of the complex of accumulated social ex-
perience and judgements which go into scientific endeavour. This
may seem a fanciful extrapolation from a simple routine test; but it
• A classic statement of the exacting skills of experimental research is provided by
Helmholtz: 'At one time, we have to study the erron of our instruments, with a view
to their diminution, or, where they cannot be removed, to compass their detrimental
in8uence; while at other times we have to watch for the moment when an organism
presents itself under circumstanca most favourable for research. Again, in the course
of our inftltiption we learn for the first time of possible erron which vitiate the result,
or perhaps merely raise a suspicion that it may be vitiated, and we find ourselves com-
pelled to begin the work anew, till every shadow of doubt is removed. And it is only
when the observer takes such a grip of the subject, so fixes all his thoughts and all his
interest upon it that he cannot separate himself &om it for weeks, for months, even for
years, cannot force himself away from it, in short, till he bas mastered every detail,
and feels assured ofall those results which must come in time, that a perfect and valuable
piece of work is done. You are all aware that in every good research, the preparation,
the secondary operations, the conuol of possible errors, and especially in the separation
of the results attainable in the time from those that cannot be attained, ccmsume far
more time than is really required to make actual observations or experiments. How
much more ingenuity and thought are expended in bringing a refractory piece of brass
or glass into subjection, than in sketching out the plan of the whole investigation I Each
of you will have experienced such impatience and over-acitement during work where
all the thoughts are directed on a narrow range of questions, the import of which to an
outsider appean triftiDg and contempbole because he does not see the end to which the
preparatory work tends. I believe lam correct in thus describing the work and mental
condition that precedes all those great results which hastened so much the development
or science after its long inaction, and gave it so powerful an influence over every phase
of human life. .•• In addition, however, to the knowledge which the student of
science acquires &om lectures and boob, he requires intelligence, which only an ample
and diligent perception can give him; he needs skill, which come only by repeated
experiment and long practice. His senses must be sharpened for certain kind ofobserva-
tion, to detect minute difFerences of form, colour, solidity, smell, etc., in the object
under examination; his hand must be equally trained to the work of the blacksmith, the
Jocbmith, and the carpenter, or the dnughtsman and the violin-player, and, when
operating with the ~ must surpass the lace-maker in delicacy of handling the
needle. Moreover, when he encounters superior destructive forces, or performs bloody
operations upon man or beast, he must possess the courage and coolness of the soldier'.
See H. von Helmholtz, 'The Aim and Progress of Physical Science' (186g), in Popu/tlr
Lenwes 011 Srierltifie SlIbjens, I . series (London and New York, 1893), pp. 319-48;
3»-1. It might be thought that Helmholtz overdramatizes somewhat; and indeed, for
routine research in safe fields, the work is not so demanding. But as we will show in
the course of this study, his list oftasb and skills is not euggerated for any genuinely
origiDal inquiry.
 Science as Craftsman's Work 83
reflects the daily practice of science, where one man will cheerfully
use data which another, working in the same field but in a different
school, will reject as nearly worthless.

Information
It is impossible to separate off the different phases of the in-
vestigation of a problem into discrete and independent units. Even
in the production of data, the later stages of work on the problem
are present in an embryonic form, as expectations on the character
of the data and its refined products, and tentative plans for the later
operations. Throughout the course of the investigation of a prob-
lem, there is a continuous re-cycling, so that the problem itself
evolves, or perhaps is desttoyed, through the interaction of the
materials which it brings into being. But there are natural points
of transition, where the satisfactory accomplishment of the tasks at
one stage makes possible the invesqnent of resources in the con-
centtated, systematic work of the next stage.
The natural successor to the production of data is its refinement
into a more reliable and useful form. For the reports on the proper-
ties of particular things and events, however sound, are still far
from being 'facts'. Their ttansformation involves a new set of craft
skills, with the application of new tools, and the making of a new set
of judgements. We can illustrate this next phase of the work, the
production of 'information', with our earlier example of readings
taken off a piece of experimental apparatus. What we do may in
practice be quite simple, but in principle it is a momentous step.
Most commonly, we plot the readings on a graph, and fit a curve to
them. This step is crucial for later work, for if we are satisfied with
the fit, we thenceforth ignore the particular readings (except per-
haps to cite them as special evidence in the published argument),
and consider the curve as the report of the properties which concern
us. Thus the data has been transformed into a new sort of material,
with the aid of certain mathematical tools. This transformation in-
volves two separate judgements, which can be assisted by special
tools, but which both involve the risk of pitfalls. The first is that
the points fit sufficiently well to a pattern, and the second is that this
particular pattern, rather than any other, is the significant one. Even
the schoolboy knows ofthe first requirement; if the points represent-
ing his readings are spattered all over the graph, then he cannot
simply draw the required curve in the likeliest place and call it a
84 The A,hievement ofS,imtiji& K1I011Jledge
day. How well is 'well enough' is a matter of judgement, frequently
assisted by statistical tools, but in the last resort depending on an
assessment of the risks and costs of errors in the particular work
at hand. 9 The judgement of whether a curve is of 'the right sort' is
more crucial and more difficult. The curve which is fitted to the
points on the graph amounts to a statement of the functional
relationship between the variables of which particular values are
recorded in the experiment. Very sophisticated methods have been
developed for assisting in this judgement, but in the last resort it
depends on the craft knowledge of the scientist, to decide which
sort of functional relation is represented by the discrete set of points
obtained from his readings. lo
If the fitted curve is of the right sort, and the points cluster
around it well enough, we may say that the information is 'reliable'.
But this is not the only test it must pass. For the statements carried

t A good CDlDpIe o( the possibility or encountering pitfalls in even the simplest

agreption or data is given in A. H. Robinson and R. D. Sale, Elemmts o/C.tovtlpiy,
yd ed. Uohn Wiley" Sons, 1969), p. 1M. The exercise is to draw'isopleths' (curves
COIUleCting domains of equal value) on a square of thirty-six unit squares; each row has
values increasing by one from the initial value, and the first column reads downwards:
123321. The isopletbs are v-chaped, and retain this shape if the unit squares are
agrepted by threes, horizontally. But if they are aggregated by threes vertically, the
isopletba become straight vertical lines I
10 The use of numerical c:alc:uIations (or curve-fitting and estimation of 'significance'
can produce numbers nearly automatically, but the need for exercising judgement is
DOt thereby obviated. For each such technique is in (act a mathematical argument,
leading to certain conclusions on the basis Q( certain assumptions on the behaviour of
the data. The appropriateness of the particular assumptions to the data at band cannot
be assumed .. given; and it bas even been argued that the 'Gaussian law of errors' is
totally inappropriate to experimental data, since it is a continuous distnDUtion and one
which gives a positive probability to errors of arbitrarily high magnitude. See N. R.
Campbell, .A" .A&tlJfllJl of tile Prineiples 01 Me"""nMlJt .u C.bltItitm (Longmans,
Green, 1928), ch. X, pp. 149-8+ He quotes an explanation for the widespread acceptance
of the Gaussian Jaw: mathematicians have accepted the law because they thought it
established by experiments, experimenters because they thought it established by
mathematics (p. 182). These considerations are not philosophical worries remote from
practice; for the numeric:al results obtained from a calculation of only moderate com-
pleDty em depend ~Iy OIl the variant of 'least squares' technique adopted. See
K. Holdea, 'The Eff'ect of RevisiODl to Data on Two Econometric Studies', The
MatMSler StluJol (March 1969), pp. 23-37. The author used two estimation procedures
(ordinary last squares, and 'limited information muimum likelihood two-mge least
squares'), on two sets of data for a fi~uation model It turned out that the para-
meters computed by the more sophisticated method were more sensitive to changes in
the cWa; and also 'The variances due to the estimation method are greater than those
due to the data revisioDs, which indicates that the choice o( estimating procedure bas
more effect OIl the parameter estimates than the choice or data' (p. 28). (I am indebted
to Dr. A. Coddington for this reference.)
 Science liS Craftsman's Work 85
forward in the work are not merely about the functional relation
implicit in the readings of the output mechanism of the equipment.
Rather, they are concerned with the properties of the classes of
things and events, of which some samples were the objects of the
experiments. These classes are inaccessible to direct view and
testing; it is only through the theory in whose terms the experiment
is conducted and interpreted, that these properties can be inferred.
It is quite possible for the experimental apparatus to produce sound
data, from which the systematic errors of the equipment have been
eliminated, and to which a good curve can be well fitted, and yet for
the interpretation of the results to be quite erroneous. For if at
some crucial point the actual workings of the experiment are
different from what is described in the theory, the properties de-
duced for the classes of things and events will be quite different
from those which actually hold. Since it is impossible to be certain
that the theory is perfectly relevant to what is going on, there can
be no certainty of the 'relevance' of the information to the work in
hand. To be satisfied that one's descriptions are really those of the
inaccessible classes ofthings and events which are being approached
through the experiment, is yet another judgement. This is even less
capable of being reduced to a routine exercise that the judgement
of reliability, and so requires an even more refined craftsman's
skill.I I
Thus the data which is first produced by an experiment must be
transformed, and tested, in order to produce genuine information;
and all these are craft operations. The situation is no different in
a descriptive science, where the data is found rather than manu-
factured. There the transition from data to information is even more
hazardous. For the data does not arise from a process which is
designed for the production of a particular effect, and which is at
least partially under control and reproducible at will. Rather, it
II There seems to be a widespread beliefthat in • matured, quantitative, experimental
science such as physics, the problems of achieving reliable information are 80 much
less severe than elsewhere that they can be considered as negligible. But the difFerence
is only one of degree, not of kind. Perhaps the most significant single experience in my
undergnduate education in physics and in the philosophy of science was an accidental
discovery of • table of values of fundamental amstants, in ]. D. Stranathan, Tile
'P.t;eles' of MotlmJ Ph,nes (Blakeston, Philadelphia, 1942), where I saw each par-
ticular quantity given several values,which differed among themselves by far more than
the 'error' assigned to any of them. I was fortunate in having been prepared for this
surprise by a previous initiation into the pitfalls of physical experiments in my labora-
tory courses; but even 80 the ground swayed slighdy under my feet.
86 The Achievement ofScientific Knowledge
reports the properties of a sample of things and events which have
been produced 'naturally', with all their variability and irrelevant
features, and where the processes of sampling and observing may
have induced further variations and distortions. 1z If the pattern to
be detected is not a simple quantitative one, then the tests for
reliability cannot use standard tools, but must rely more heavily on
personal judgement. Similarly, the judgements of relevance are even
more hazardous than in the case of experimental fields; for the lack
of control over the materials described in the data, and the inferior
precision and elaboration of the arguments, inevitably reduce the
rigour of the testing of the relevance of any piece of proffered in-
formation. The distinction between reliability and relevance in
descriptive fields can be seen from the example of the testing of
individuals for psychological and educational purposes. One set of
skills and controls are involved in consttueting tests that are 'reliable',
in that a subject will score nearly the same aggregate marks every
time. But what is called the 'validity' of the test, the relevance of
the numerical score to the attribute which is the object of assess-
ment (be it 'intelligence', aptitude, or even mastery of a particular
academic subject) is quite another matter.
The category of 'information', as the refined product of the data,
can also be useful in the analysis of purely theoretical or mathe-
matical sciences. For this corresponds to the set of arguments which
are sufficiently elaborated to be recognizable as components of a
completed proof. They must still be checked for their internal
12 The hazards of the transition from data to information in descriptive sciences is
shown particularly wen by what might be called the Case ofthe Bulgarian Pigs, described
in O. Morgenstern, 0. tile Altl6lUY ofEe01UJflli& ObsmJtltitnu, 1st 00. (Oxford University
Press and Princeton University Press, 1950), p. 19. The data are the number of pigs in
Bulgaria on 1 January 1910 and 1920; these were 527,311 and 1,089,699 respectively.
The information derived from this pair of figures WclS that the number had more than
doubled during the decade. Unfortunately, the inference was faUacious; what had
happened in the interim was that the calendar had been changed from the Julian to the
Gregorian, 80 that the peasants c:elebrated Christmas in January (new style) just as the
Soviet Union celebrates the 'October Revolution' in November. Just half the pig
population is slaughtered immediately before Christmas; and 80 the extra pigs on 1
January 1920 were merely awaiting their yuletide extinction.
Similar pitfalls beset the more sophisticated transition from data to information in
history. In 'The Elizabethan Aristocracy, an Anatomy Anatomized', Eeonomie History
RevieJ" 2nd series, 3 (1951), 279-98, H. R. Trevor-Roper discusses the use of the sums
of money quoted in documents known as 'bonds', as information about the size of a
loan. He argues that such documents were far more likely to be a record of the seturity
for the loan, which would be at last twice as great as the loan itself: The conclusions
about the 'impoverishment' of the Elizabethan aristocracy are modified accordingly.
 Science as Craftsman's Work 87
coherence, lest there be hidden flaws in them; this corresponds to
the reliability of the information in sciences based on experience.
Also, since each argument establishes a property ofa particular class
of conceptual objects, then the mutual relevance of the different
arguments must be checked, lest the overlap between the classes
discussed in the various arguments be too small to provide a
significant result. The set of such arguments does not yet constitute
a proof, for the different components must still be welded together
into a coherent argument, and in the course of this work their in-
ternal reliability and mutual relevance must be tested again.
Even in fields based on experience, the craft work of achieving
scientific knowledge is not completed with the production of in-
formation. For the properties of the objects of scientific inquiry
are not fully established by a mere list of reports. Rather, an argu-
ment of some degree of elaboration is necessary for conclusions to
be drawn about these objects. In this argument the information is
put to use as evidence; and for this function it is subjected to
further refinement and testing. The elaboration of the argument is
of course a matter ofdegree; in some descriptive fields, the necessary
argument may be quite rudimentary, and a collection of information
can constitute a worthwhile piece of work. I shall discuss these later
phases of the investigation of scientific problems in the next chapter,
whose theme will be the artificiality of the objects of scientific in-
quiry. For the remainder of this chapter, I shall analyse some other
aspects ofscientific work where its craft character is clearly exhibited,
and is of importance for an understanding of the social aspects of
the work.
Only a part of the information used in a problem will derive from
the data produced in its investigation. The rest is taken over from
other work, either from individual research reports or from collec-
tions of standard materials. I 3 Although a particular bit of this
material may be generally accepted as solid fact, it must still be
considered as no more than possible information; for its reliability
13 The proper handling of the stock of 'existing information' relevant to • problem is
a craft skill whose most elementary part is 'information retrieval' techniques. On this
task, the scientist has no substitute for his judgement and experience, since the estab-
lishment of those classification categories which will produce 'relevant' materials, the
decisions on where to search and where not to search, and the quick assessment of the
quality of the materials examined, are all very .i'.btlc tub. A common pitfall is to
require 'all' the information before engaging on the work, to avoid errors and duplica-
tion; this is precisely analogous to designing an aeroplane that cannot cruh-one that
cannot fly.
88 The A,hit'Dtment ofScientific Knowledge
and relevance to the work in hand can never be taken for granted.
If this 'existing information' relates to things which are remote
from the problem in which it is used, then it must generally be
accompanied by an explanation. This will not be designed to impart
a deep understanding of the material, but only to enable the user
to handle it competendy and to avoid the most common pitfalls.
Of the stock of established facts in a field, only a small part of that
which survives in general accounts and teaching are results in-
teresting in themselves. Generally, a fact is preserved from oblivion
only when it is useful as existing information, performing a function
in new work analogous to that of a tool. We shall later see that the
patterns of evolution of established facts are conditioned by this
function; and indeed give rise to the special properties of per-
manent scientific knowledge.

Tools
In the production of experimental data, the apparatus is easily
recognized as a special sort of tool, similar in its functions to the
tools of a handicraftsman. They are not the objects of the work, but
are the means by which those objects are created and shaped. The
tools of the scientist are not all physical equip.ment; in the refine-
ment of data into information, other tools are brought into use,
most noticeably statistics and other mathematical representations.
The physical and intellectual tools of scientific work vary in their
complexity and sophistication; some require no special skill or
training for their use, while for others the scientist must have a
craftsman's knowledge of their powers, limits and possible pitfalls.
Every field has its own special tools, of which some require an ex-
tensive technical knowledge, supplemented by craft experience, for
their use. It is for this reason that a lengthy apprenticeship is neces-
sary before anyone can embark on independent work in a developed
scientific discipline. We shall see how tools condition the work which
is done in any field of science, and also enter into the social relations
between different disciplines.
We can make a rough classification of the different sorts of
materials which can be considered as tools, starting with the physical
apparatus by which data is produced. Following on this are the tools
which actually transform data into the shape appropriate for in-
formation and test it. Such data-processing tools can be statistical
techniques, or machinery which is programmed to exhibit patterns.
 Science as Craftsman's Wori 89
There is a third class of tools, which do not actually produce or
transform data, but which make possible the interpretation of the
data or of the information. One example of this class is a corpus of
standard information about the objects of the inquiry, assembled
in handbook form. In sciences dealing with large collections of
different objects, noticeably the descriptive sciences, the need for
such tools is particularly strong, and there will develop entire
sub-disciplines devoted to their production. From this corpus of
standard information, materials may also be taken for inclusion as
evidence (after the appropriate tests) in the argument of the prob-
lem. In this same class are 'tool-subjects', other fields of science
which must be mastered to some extent in order that competent
work can be done in the given field. The relation of being a tool-
subject to another field, which links many pairs of fields, is in
general asymmetric; those fields which deal with the more abstract
and general properties of matter function as tool-subjects to those
less so. This asymmetry is of great significance for the development
of science over the last few centuries, and will be examined more
closely.
Finally, there is another important class of tools, of a significandy
different character. These are required when the language of the
argument of problems in a field is not merely the vernacular en-
riched by technical terms, but when it is a formal, artificial language
of its own. The most common example of this is the use of the
objects of some fields of mathematics (as the calculus) in theoretical
sciences. Such fonnallanguages are not merely a ttanslation of the
vernacular. They provide power to the arguments, making possible
inferences which could never be made in the vernacular; and also
impose a precision (not necessarily accuracy) on the objects of the
arguments. Also, the potentialities of the formal language for the
construction of arguments are revealed by continuing research in
the formal system as an object in its own right by specialists; and
the avoidance of pitfalls in its use is learned partly by craft ex-
perience but also by deeper knowledge of the mathematical system.
Hence the scientist who uses some branch of mathematics as a
language-tool must render himself competent in its practice and
theory.
Although tools are auxiliary to the advancement of scientific
knowledge, their influence on the directions of work done is im-
portant and frequendy decisive. New tools make possible the
90 The Achievement ofScientific K"lIJl)ledge
production ofentirely new sorts ofdata and information. As the tools
of a field develop, projects which previously required outstanding
talent or great perseverance, now become routine and nearly trivial.
This effect can be seen even by students, who may first study
physics without the calculus, and later discover how the laborious
and roundabout methods of describing and deriving various effects
are rendered unnecessary by this powerful tool of argument. Also,
in fields where the data comes from experience, it must be produced
in a certain general form, so that the appropriate tools can be applied
to it. This is most obvious in the case of statistical tools; many are
the cases where a consulting statistician throws back a set of data
to a surprised and somewhat hurt scientist, telling him to discard
that lot and bring back another set of a better statistical design.
In this requirement on the data we see the first of the reasons why
one cannot simply go out and get data as an independent foundation
for later work. Finally, each set of tools has its characteristic pitfalls,
and if blunders and disasters are to be avoided, the user must de-
velop a craftsman's knowledge of their properties. The abuses of
statistics are again the most notorious example of this;'. but even
in descriptive sciences, the uncritical use of handbook information,
which is always incomplete, obsolescent and not quite fitted to the
needs of the work at hand, can lead to the most astonishing blunders.
A consideration of tools also leads us directly into some important
social aspects of scientific activity. Because so many of the essential
tools for any field of science are so highly sophisticated, to achieve
complete mastery in the use of some of them involves becoming a
specialist in the tool rather than in the field to which it is being
applied. There is thus a natural division of labour between tool-
experts and their clients; and the tool-experts are not merely
individuals serving as auxiliaries to the clients in the work, but
themselves can form a self-contained speciality, a tool-providing
field. When two craftsmen with different skills are involved in the
same project, they will inevitably see the work from different
points of view. The different approaches will be complementary,
and can correct and enrich each other; but they can also be the occa-
sion of conflict. For each of the partners may be wanting to get
something slighdy but significantly different from the project. The
client wants data or information, reliable by the accepted standards
14 The classic primer in the abuse of statistics is Darrell Huff, HOJlJ to Lie Jl'it"
StlJlisti&s (Norton, New York, 1954; GoUana, 1958).
 Science as Craftsman's Wori
91
of adequacy of his field, with a minimum of expense and delay;
while the tool-expert, unless he is completely subservient, will be
looking for opportunities for developing particular tools in which he
is interested The two parties even perceive the situation differendy,
for their different interests correspond to different bodies of craft
knowledge; and unless both parties enter the relation with consider-
able mutual comprehension and respect, only their respective incom-
petences will be communicated, and conflict will ensue. How such
conflicts work out depend on the social relations, and relative
strengths, of the clients and the tool-experts, individually and as
recognized collectives. It must not be thought that the tool-experts'
function as auxiliaries keeps them in an inferior position; they are
not merely essential for the work, but they may frequendy have
command oftechnicalities which are incomprehensible to the clients,
and which also bear prestige in themselves. This is most noticeable
where the tools are mathematical in character. IS
Not all tools are so specialized as to require a ~istinet corps of
experts for their use. In such cases we may speak of tool-users, not
15 The development of electtonic computers could be the subject of a useful case-
study in the history of modem technology, if attention is paid to the different purposes
of the users, the experts, and also the manufacturers. Although the development of the
operating characteristics of these machines can be seen as a steady improvement with
occasional forward leaps, their history from the users' point of view can be seen as a
series of partially resolved crises. On this, see C. Strachey, 'Systems Analysis and
Programming', Scimtiji& AmericlIn, 21S, NO.3 (September 1966), 112-26. The diverg-
ence between the purposes ofusers and experts can be seen in the criterion of 'efficiency'
of a computer program; whether it is assessed by the proportion of 'computing' time
in the total length of a program, or by the convenience of the user. Inexpert users do
not constitute a coherent group, and so their convenience does not rank high among the
specification of the functions of the computer. Only occasionally can they even publish
their criticisms, since to do so requires a level of expertise which resolves the individual
difficulty. One such publication is M. Hamilton, 'Computers for the Medical Man: a
Solution', British Metlicill JOUNUI1, 2 (30 October 1965), 1048-50. The author had found
that the so-called 'standard' programs for simple data processing are nothing of the
sort; and succeeded in devising some which could be used and interchanged without
extensive re-writing every time. The deficiencies of computers from the users' point of
view will be discussed quite frankly by manufacturers on some occasions, but these will
tend to be in the context of advertising an improvement which obviates the previous
difficulties. Thus Hewlett-Packard describes the prevailing situation asjollows: 'Given
the chance, the computer can live up to its promise. But in all too many labontories,
the computer doesn't even stand the chance of a trial because it aeates new problems
that some scientists consider to be worse than the old. ouer among these is the com-
plexity of putting the computer to work in the laboratory-programming it, mastering
the instrument-computer and the man-machine interfaces-which, to the scientist, is
often a greater drudgery than the manual data gathering and calculation that the com-
puter eliminates.' See Science, 167 (1970), 1196-7. They offer small, instrument-
oriented digital computers, and also shared time with packaged programs.
92 The Achievement ofScientific KnoJlJledge
merely clients, and of tool-using fields in relation to these products
of the tool-providing fields. There is a general asymmetry between
tool-providing and tool-using fields, which, when combined with the
prestige-ranking among sciences that we have inherited from the
past, produces a systematic tendency for tool-providing fields to
assimilate those which they nominally serve. The relation is a simple
one: those fields which describe the more abstract and general
properties of matter serve as tool-providers for those which are
concerned with the more concrete, and more particular or more
highly organized. The sequence runs from mathematics to physics,
chemistry and biology; attempts to tack the human sciences neady
on to the end of the scale have not as yet proved successful. The
reasons for this asymmetry are twofold. First, the tools which are
designed for the production of data or information in the more
particular and organized properties of matter, must necessarily be
designed to cope with complexity and variability which is en-
countered in such situations, rather than to attain the precision
which is possible in simpler and more stable contexts. Hence tools
designed for biological work will find no use in physics. On the
other hand, some of the more abstract and general properties of
matter may well be relevant to a problem in a field such as biology,
especially in the production and testing of data and information;
we have seen this in the case of statistics. This asymmetry is re-
flected in the teaching of the various sciences; the physicist's tool-
subjects are mathematics and certain techniques of instruments,
while the biologist must learn something ofall those sciences further
up the scale.
This asymmetry of provision and use does not imply that tools
are created within the confines of the more abstract disciplines, and
then hired out to those lower down on the scale. For it will fre-
quendy happen that a tool is brought into being to perform a func-
tion in a problem in a 'complex' field, is then seen to be capable of
extension in all directions, and may even gready influence the
development of the tool-providing field itself. The history of
statistics, many of whose most powerful techniques were developed
in just this way, is a case in point; and many (although not all)
branches of mathematics have evolved directly out of tools designed
for special functions in the natural sciences. 16
16 The advances in statistical "technique associated with such names as Galton and
R.. A. Fisher were achieved in the aeation of tools for analysis of numerical data in
 Science as Craftsman's Work 93
We have already seen that in a general way the tools which are
available define the range of problems that can be studied. But the
influence of tools on a field can be more subde than a mere creation
of possibilities. The extensive use of a tool involves shaping the
work around its distinctive strengths and limitations; one can rarely
apply a new tool to an existing stream ofresearch without modifying
the stream strongly. Hence as new tools come into being, and are
judged appropriate and valuable by people in the field, they alter
the direction of work in the field and the conception of the field
itself. The men of an older generation who cannot master the new
tools may grumble that the field has been distorted or taken over by
outsiders. Whether they are right is a matter that only a later history
can judge. But any such judgement must be made in terms of an
implicit philosophy of science; and in the one which has been
dominant for several generations, a field becomes more genuinely
'scientific' as' it more closely resembles theoretical physics. Hence
the natural tendency of the available tools to modify a field in that
general direction are reinforced by the prestige of the limiting point
of the process. It is important to realize that this confluence of the
technical aspects of tools, and the prevailing metaphysic, is an
historical accident rather than an essential feature of natural
science. If the 'paradigm' natural science were to become a discip-
line like ecology, which uses the whole range of tool-providing
sciences but whose objects cannot be reduced to those of any of
them, then the social relations of tool-providing and tool-using
fields would be drastically altered.
The combination of these two asymmetries, one inherent in the
work, and the other socially imposed, has produced a false picture
of the relations of the different scientific disciplines. It is commonly
supposed that physics is not only stronger than biology as a field,
but also more fundamental in that its problems and objects are
independent of the latter. When two such fields are connected, as
in biophysics, the penetration is all in one direction. Yet this is
not necessarily so, for in the past it has been otherwise, on most
significant occasions. For example, few physicists will be aware that
the conception of the conservation ofenergy, which ttansfonned the
the biological and human sciences; they soon generated lively fields ofpurely theoretical
research, cast in a form comprehensible only to mathematicians. In mathematics, the
field of differential equations was origiDally a tool for rational mechanics; and Fourier's
series was invented and developed in the course of investiptiops into the theory of the
flow of heat in solid bodies.
94 The Achievement ofScientific Kno1lJ/edge
physical sciences in the middle of the nineteenth century, arose
partly out of biological problems which were investigated in a
consciously philosophical style. 17 The full history of the discovery
of the First Law of Thermodynamics shows that this surprising
origin was no 'accident'; but to explain it lies beyond the com-
petence of the traditional philosophy of science, and the physicist's
common sense of science on which it is based.
Finally, this discussion of tools shows how subde must be
the craft knowledge of a scientist who is a leader in his field. He
must not only be able to develop a craftsman's competence in the
use of particular tools, and have a sense of their powers and limita-
tions; but, if he is to do anything but follow fashions set by others,
he must assess the sorts of problems into which the use of par-
ticular tools would lead him and his colleagues and decide whether
they are appropriate for the best progress of work in his field.

Pitfalls
The craft character of scientific work is exhibited most system-
atically through the concept of 'pitfall', which I have already men-
tioned several times. IS The importance of this concept can be seen
from Francis Bacon's aphorism, 'What in observation is loose and
vague, is in information deceptive and tteacherous.' Leaving aside
his distinction ofobservation and information, we may ask why there
should be deception and treachery in the inference? This is very
strong language indeed. What does it mean? Whenever we extend
from the known to the unknown, we do so on the basis of expecta-
tions of what we will find; these are necessary to give direction to
our moves, and to supply interpretations of what we encounter.
These expectations are always incorrect in some measure, and so
we learn through the discovery of our errors. However, not all our
errors are so considerate as to announce themselves as soon as they
are made. It can happen that we follow an erroneous path of in-
17 For the importance of NlI'wpllilosoplUe in the inte1lectua1 background of many of
the 'discoverers' of amservation of energy, see T. Kuhn, 'Energy Conservation as an
Example of Simultaneous Discovery', in Criti&1Il Problems i,. ,Ite History of SeieMe,
eel. M. Clagett (University of Wisconsin Press, 1959), pp. 321-56, especially pp. 338-9.
The case of Helmholtz has been studied in great depth by Y. EIkana, 'The Emergence
of the Energy Concept' (ph.D. Thesis, Brandeis University, 1967).
18 I learned the concept of 'pitfall' as part of the aUt wisdom of mathematics from
my teacher Professor A. S. Besicovitch, F.R.S. He had the aphorism, 'In analysis, the
pit&lls are everywhere dense'; its full flavour can be appreciated only by those who
know the technical meaning of the latter tent1.
 Science as Craftsman's Work 9S
vestigation for some time, investing great resources into its pursuit,
and only much later discover that we are mistaken. Then we realize
that our time and labour had been wasted, and perhaps our con-
fidence and reputation shaken as well. Bacon knew that we cannot
but try to confirm our expectations as we advance into the un-
known; and this is much easier if the things we find, as we go along,
can easily be interpreted in our favour. In this fashion, with 'loose
and vague' materials, we can deceive and bettay ourselves, to our
eventual undoing.
Thus the path of discovery is beset by concealed ttaps for the
unwary, which we can call 'pitfalls'. In some ways these are more
dangerous than the physical hazards from which they take their
name, for one learns only in retrospect that one has stumbled into
a pitfall at some earlier point of the work. One may produce data
on equipment which has a concealed systematic error; or the theory
of the equipment may have a hidden error which vitiates the in-
ferences from the readings; or the first sketch of an argument may
have an ambiguity or false deduction which undermines all the
reasonings established on its basis. At every stage of our exploration
of the unknown, we are at risk of being mistaken, and of remaining
in ignorance of our mistakes until irretrievable damage has been
done. 19
The encountering of pitfalls through the making of judgements
influenced by expectations is shown most clearly in the hazards of
19 Pitfalls in observation are most likely to occur when the 'data' are subjectively-
assessed extreme or nul points of a visual phenomenon. This was responsible for the
short-lived inquiry into 'N-Rays' early in this century; see D. J. Price, Samee si1l&e
Babylon (Yale University Press, 19S7), ch. 4- The identical pitfall vitiated the 'Allison
effect', whereby the clements 'Alabamine' and 'Virginium' were 'discovered'. See
F. G. Slack, 'The Magneto-Optical Method of Chemical Analysis', Jowul of tM
F'lInltlin Institute, 218 (1934), 44S-62. (I am indebted to Mrs. R. Countryman for this
reference.) It is interesting that the 'findings' in this second case were longer-lived;
even after the war American students of chemistry learned of these two elements by
their pattiotic namcs, rather than as Francium and Astatine.
Q!tantitative experiments can also be subject to disastrous pitfalls through the
scientist's ignorance of distorting effects on his equipment. The experiments ofD. C.
Miller with an improved interferometer, by which he detected an absolute motion of the
earth through the aether from 1925 onwards, were probably vitiated by temperature
effects, in spite of the warning on that point by Helmholtz to Michelson and Morley in
the 18805. See L. S. Swenson, Jr., 'The Michelson-Morley-Miller Experiments before
and after I 90S', JOUNUII oftM History ofAstronomy, I (1970),68. Temperature effects
also produced a pitfall in an experiment of otherwise classic simplicity, testing whether
photons lose or gain energy when moving against or with • gravitational field. In this
case, the pitfall was exposed by a Cambridge (Trinity College) undergraduate, who has
since gone on to a distinguished career in physics. See B. D. Josephson, 'Temperature-
 96 The Achievement ofScientific Kno1lJledge
using research assistants, working essentially as technicians, for the
production of data. For the assistant will generally know what his
supervisor expects to find; indeed, he must have such explicit ex-
pectations if he is to make the first judgements on the soundness of
the data. But when unexpected and contrary results appear, he
must make a judgement on their significance, balancing his own
limited technical competence against the superior understanding of
his master, and perhaps being influenced by political considerations
as well. The natural course of action is to present information from
which the anomalous data have been expunged. If the supervisor is
concerned to have evidence derived from genuinely sound data, he
must go to the ttouble ofttaining his research assistants in genuinely
critical and independent thinking as well as in the craft of the
research. Sometimes this may be more trouble, and still not so
reliable, as doing the work oneself: This phenomenon indicates that
there may be an upper limit to the degree of division of labour, in
the dilution of the sophisticated craft skills, in worthwhile scientific
research. %0

Dependent Shift of Rays Emitted by a Solid', Phyli&lIlRtTJitJl1 Letters, 4 (1960),341-2.

(I am indebted to Professor F. A. E. Pirani for this reference.)
The possibility of pit&lls being discovered in well established experimental fields is
shown by a report, 'A Cautionary Tale for Chemists', NnIJ Scientist, 4S (1970 ), 543.
Two c:hemis1s, in the course of a straightforward kinetics study, discovered that an im-
portant component of one of their reactions was the glass wall of the flask containing it;
the glass was not inert against dilute sodium hydroxide. The article concludes: 'The
studies indicate the need to interpret more carefully than has hitherto been the case
all chemical data measured in this type of system. Chemistry, it seems, is still capable
of springing a few surprises.'
Very few scientific fields have standard literature on the pitfalls of their characteristic
patterns of argument. In quantitative social science, the classic is W. S. Robinson,
'Ecological Correlations and the Behaviour of Individuals', American Soaolo,;clIl
Rmnz" 15 (1950), 351-7. This is discussed and further developed in E. Allardt, 'Aggre-
pte Analysis: the Problem of its Informative Value', in 'QUllnt;tllt;vt Ecolo';&111
Aulysis ;,. tM Socilll Samees, eels. M. Dogan and S. Rokkan (M.lT. Press, 196c).
20 I recall a colleague te11ing me of his admintion for a particular foreign research
student who, although DOt brilliant, did overcome his natural humility sufficiendy to
inform the supervisor that the data were DOt as expected. The student had been so
conditioned to believe everything said by a Professor in his own country, that it required
considerable courage for him to make the criticism. At this very fundamental level,
there is clearly some correlation between independence of thought and worthwhile
scientific work. A similar talc with a somewhat different ending is told by Lincoln
Steffens in his AlllolJioVllphy (Harcourt, Brace, New York, 1931). While studying under
Wundt in Leipzig he and his colleagues came across the laboratory records of another
American who had been there before. From them they could reconsttuet the history
of his research, where he produced data 'which would have given aid and comfort to
the enemy and confounded one of Wundt's most axiomatic premises'. This would also
 Science as Craftsman's Work 97
From the nature of the work, it is impossible to eliminate pitfalls
from scientific inquiry. The experience of matured scientific dis-
ciplines is that they can largely be avoided. This is done in two ways:
by the charting of standard paths which skirt them, and by each
investigator becoming sensitive to the clues which indicate the
presence of the special sorts of pitfalls he is likely to encounter in
his own work. The first of these requires a tradition of successful
work in the subject where there is a body of standard techniques
which can successfully be applied as a routine. When these are
taught to students only some of the pitfalls which they have been
designed to avoid are pointed out. But these techniques are an
embodiment of successful craft practice, built up over generations.
Then, when an individual scientist explores beyond the range of
the well-established techniques, his craft knowledge must neces-
sarily be more subde and personal, for the pitfalls he is likely to
encounter are peculiar to. his particular materials and tools. The
clues to the presence of pitfalls are all he has, since he is beyond the
range of the charted paths. Thus the accumulated social experience
of his field must be supplemented by his personal craft experience
of his portion of it.
This discussion of pitfalls is of some philosophical significance,
for it indicates one of the reasons why it is vain to seek for certainty,
or even for a guarantee of the existence of certain knowledge, in
science. There can be no absolutely certain foundations in experi-
ence, nor any absolutely certain inferences from that experience, for
the achievement of knowledge of the natural world. Worse still,
from this discussion of pitfalls we see that we even lack a guarantee
that our errors will be revealed to us by a direct and straightforward
process. It is possible to argue from the impossibility of complete
certainty in science, to a 'sceptical' position, asserting that all our
supposed scientific knowledge is merely hypothetical, or probable,
or illusory. But such abstract and general arguments do not indicate
why some of our knowledge is more certain, and more solid, while
some is less so. The concept of 'pitfall' furnishes one criterion
whereby we can make such a distinction. In an established discip-
line, a trained man can work reliably in the realm of the known,
have been damaging to a career full of promise; so the student solved the problem by
'changing the figures item by item, experiment by experiment, so as to make the curve
of his averages come out for instead of against our school'. Full of admintion for this
mathematical feat, Steffens none the less concluded that experimental psychology did
not provide a foundation for a science of ethics, and turned elsewhere (pp. 150-1).
98 The Achievement ofScientific Knowledge
and successfully feel his way forward into the unknown. But this
happy state of affairs does not always hold; and indeed every one
of our firmly established sciences has a long prehistory of its im-
mature state. Then, even the accepted knowledge was liable to be
revealed as illusory, and in the vanguard of the science, each man
groped forward blindly, hardly ever succeeding in avoiding the
pitfalls in his way. We can see that the establishment of a successful
discipline requires a man of exceptional talent and courage, for he
must not only succeed in his researches where all others have failed,
but he must create the craft knowledge of the pitfalls of the inquiry,
so that the lesser men who follow him can safely proceed to the
development of his own achievements. 21
Thus we can distinguish between more and less solidly established
knowledge, by the degree to which the ways around its common
pitfalls are well charted, and those encountered in the extensions of
that knowledge can be sensed in advance. This criterion is not
sufficient in itself, for the progress of science renders most of its
earlier achievements irrelevant and obsolete. However, the con-
cealed imperfections which are present in scientific knowledge at
any point of its development, are not the same as pitfalls. A pitfall
is the sort of error that destroys the solution of a problem, and
nullifies its conclusions about the objects of the inquiry. We do not
say that Newton's assertion of the existence of absolute space and
time, or Boyle's ignorance of van der Waals' equation, were 'pit-
falls'. Although our knowledge is more refined, we do not contest
the solidity of their achievements. Of course, these are extreme
examples, and we cannot draw a perfectly sharp demarcation
between perfectible knowledge and results vitiated by pitfalls. But
as a matter of historical experience in a matured discipline the line
al This progression can be discerned in the decades of work that were required for
GaIileo'. lDIItery of the principles of bodies in non-unifonn motion. His first formula-
tion of the law of acceleration of falling bodies was the more common one of his time,
that speed increases in proportion to distance traversed. He actually succeeded in
'deriving' from this false relation the result already known to him, that distance is
proportional to the square of time. He eventually discovered the falsity of this rule: I
think it likely that he found it impossible to relate it to other known results of his (such
as, equal times of descent along all chords of a circle), or to base a theory of projectile
motion on it. When he finally came to disprove it, one of his two arguments was itself
fallacious I See Galileo, DisetWsi •••, translated u D;"lopes COII&tmi", t1IHJ tJnI' SeinJ&es,
tr. H. Crew and H. de SIlvio (Macmillan, New York, 1914; reprint Dover, New York,
DO date), pp. 167-8 (Edizione Nazicmale, viii, 203-4). For a discussion of the faIsc rule,
lee A. R.. Hall, 'GaIiIeo'. Fallacy', Isis, 49 (1958), 343-4. followed by discussion by I. B.
Cohen and S. Drake.
 Science as Craftsman's Work 99
is fairly sharp; and this is 1nother expression of the fact that solid
knowledge derives from a tradition of successful craft work in its
achievement. 22
At this point of my discussion, as at others to come, I must con-
sider the question of why this important feature of science has
hitherto escaped notice. It is easy to see why all the earlier ttaditions
in the philosophy of science ptainly concerned with epistemological
problems should have ignored the craft character of scientific work,
and with it the concept of pitfall. In several of the human sciences,
the craft character of the work of advancing knowledge has been
recognized by leading men, and through their writings is familiar to
all those who are concerned to understand their own work. 23 Hence
unless we believe that the study of the natural world is
utterly different, in its work as well as in the character of its
conclusions, my thesis will become obvious as soon as it is
realized to be significant through the changing common sense of
sCIence.
I must also reckon with the historical fact that the formal training
of scientists has generally been carried on without any recognition
of the craft character of scientific work. There are several reasons
for this, of which one is the implicit philosophy of science which has
prevailed up to now. Also, most of the basic knowledge which a
scientist will need to have available as tools can be organized into
a systematic form in which the pitfalls are bypassed without any
immediate need for their being identified. For the pitfalls are en-
countered only when this knowledge is put to use, and they depend
so much on the particular application, and are so various, that a
comprehensive discussion of them would be quite impossible. In
laboratory courses, students are given a gende introduction to the
pitfalls likely to be encountered in the use of physical tools for the
production of data, and they learn the craft techniques for manipu-
lating these tools reliably. However, this essential aspect of labora-
22 I am grateful to Mr. Keith Boughey, then an undergraduate at the University of
Leeds, for raising the problem solved in this paragraph during a discussion in 1967.
23 Among practising historians who have written on 'historical method', the aUt
character of historical research is either recognized explicidy or is implicit in the whole
argument. An illuminating recent exposition is G. R. Elton, TAe PrlJ&ti&e of History
(Collins (The Fontana Libary), Sydney University Press; London and Glasgow, 1967).
In c. Wright Mills, TAe Sodologi&1I1 It1U1,;""titm (Oxford University Press, 1959; The
Grove Press, New York, 1961), the idea of craftsmanship runs through the book, and
an appendix 'On Intellectual Cnftsmanship' offers an account of the author's own
style.
100 The Achievement ofScientific Knowledge
tory teaching is frequendy ignored, and students usually believe
that they are merely 'verifying' for themselves that certain standard
effects can be reproduced.
A recognition and systematic use of the phenomenon of pitfalls
might be very effective in the teaching of those simple but essential
craft skills which are involved in scientific, scholarly or administta-
rive work. A person trained for conceptual thinking generally finds
the books on 'method' in such fields to be of little use; there seems
to be little fruitful middle ground between boring recipes and mean-
ingless aphorisms. But an exposition of standard techniques in
terms of the pitfalls they are designed to circumvent, with examples,
could go far to make them meaningful and obviously worth master-
ing. 24
It is only in the training of research students in science that the
craft character of scientific work is now explicidy recognized, and
with it the importance of learning how to sense the presence of pit-
falls. Everyone who supervises research students knows of the
dangers of giving too much, or too little, help and guidance. If the
student is simply told of a problem-situation and then left to his
own devices, he may spend so many months floundering among
failed attempts even to set a problem that he becomes completely
discouraged and gives up before he has produced anything solid.
But if he is given constant advice, he proceeds all too smoothly
through his project, being little more than a reliable pair of hands
for executing the supervisor's ideas; and he never learns to grapple
with his materials or to sense the pitfalls occuning in his field. How
much help should be given depends on the project, the student, the
policy of the school, and the attitude of the supervisor. Research
supervision is itself a craft, the most subde and demanding sort of
teaching. Some research schools let their students stumble into a
large proportion of the pitfalls in their way; this is expensive, time-
consuming, wasteful of talent and hard on the students. But those
who survive such a course are skilled craftsmen, and of tested
sttength of character in this respect as well. In others, research
students are given plenty ofguidance, so they are sure to get through
their work in a reasonable time, without too much suffering.
Wastage and agony are thereby reduced, and 'success' is more
likely, especially if the research student is really functioning as
M This has been done for architeelB, in H. B. Cresswell, Tile HflMYllJootl File (London,
1929; reprinted, Faber & Faber, London, 1943).
 Science as Craftsman's Work 101

low-paid skilled manpower working on the supervisor's problems. 25

But those who emerge from such an experience with a Ph.D. as a
formal certificate of competence, are as yet unskilled and untested
in original scientific research. The one extreme policy results in
many failures, and in a few scientists of promise; the other produces
competent 'scientific manpower' on a production-line basis. The
character of the scientific work done by the graduates of the differ-
ent sorts ofresearch schools will inevitably reflect their training; and
here we see one way in which the craft work of science is influenced
by the institutional and social context in which it is conducted. 26

Craft Methods: TechnitJues

We have already seen several ways in which the work of scien-
tific inquiry requires knowledge which is learned only through pre-
cept and experience in a multitude of particular cases, and which
therefore is not 'scientific' in character. The assessment of data and
of information, and the manipulation of tools, are all subject to
pitfalls; and it is only the craft knowledge of the investigator which
enables him to avoid some and sense the presence of those which
remain. Such knowledge forms part of the corpus of methods, those
principles and precepts which guide and control the work being
done. For the present, we can further illuminate the craft nature of
scientific inquiry by discussing those methods which are most
closely related to the particular tasks undertaken, and which there-
fore consist more of completely tacit knowledge and particular pre-
cepts. Those methods which condition the assessments made of
problems, and are more concerned with their function as social
possessions, will be discussed systematically later on.
25 Thus E. Rudd, 'Rate of Economic Growth, Technology and the Ph.D.', Mi"",.,
6 (1g68), 366-87; 'When we interviewed CUI'I'alt research students in chemistry depart-
ments and asked them if their research topic was closely related to that of their super-
visor, sometimes the reply was: "What do you mean, closely related to that of my
supervisor? It is my supervisor's research.'" (p. 382).
26 Because research supervision involves the traDsmission of the most subtle skills,
it may give a clue to one of the causal factors involved in the 'direct square law of
association' known in the folklore of science. This is, that men of the first rank will
associate with men of the first rank; those of the second, with those of the fourth, etc.
For a scientist of moderate ability and commitment will generally Dot be actively
concerned about the craftsmanship mvolved in his work, and even less able to impart
the skills and the spirit of craftsmaDship to the level that he possesses. Hence the craft
skills of research in a field can become attaluated very npidly in transmipion through
a sequence of lesser men.
102 The Achievement ofScientific Kno1lJledge
Of all the 'methods' ofscientific inquiry, the most straightforward
are the techniques of routine manipulations of tools. This is usually
the content of courses in 'methods' where materials from a tool-
providing field are made available to potential tool-users. The tools
designed for such standardized uses are usually made as close to
'foolproof' as possible; so that by following the simple rules the
user should be able to avoid pitfalls in any ordinary applications.
However, perfection in this as in anything else is not easily achieved,
and so the simple rules of manipulation may be supplemented by
somewhat more subde and particular instructions and precepts for
avoiding pitfalls which might produce obvious damage (as to the
tool itself and its user) or hidden errors of interpretation of
the products of the operation. These too can be classed as
methods; they are the first crude, explicit elements of craft
skill in using the tool, as distinct from merely mechanical
manipulation.
In general, the contents of such elementary points of technique
will be precepts rather than principles. For usually the tool-user
cannot be in sufficient command of the theory of the tool to appre-
ciate the explanations for its operations, or of its malfunctioning,
even when such explanations exist. As an elementary example of
how a 'method' is invoked for avoiding such pitfalls, the beginning
student in chemistry has no basis for knowing that concentrated
sulphuric acid evolves heat on mixture with water, could not under-
stand the explanation if it were given to him, and should not have
to learn the phenomenon by 'trial and error'. Hence he will simply
receive a precept to drop the acid gradually into the water, rather
than to mix all the masses or to drop the water into the acid; and
the 'explanation' of the precept is restricted to an assertion of the
dangers of the other courses.
For a full mastery of the use of tools, explicit precepts are in-
sufficient. Any extension of the uses of a tool involves new hazards
ofpitfalls; and the unknown cannot be described by formulas. Hence
the full craft knowledge of any particular tool, for a particular range
of functions, will involve a large measure of personal experience. To
the extent that the personal knowledge of a tool is deep and subde,
any set of explicit precepts will fall short of conveying it. In any
real situation there are too many subde cues, and too many partly
relevant precedents, for the knowledge of how to cope with novelty
to be reduced to tables of experiences and inferences. This aspect
 Science as Craftsman's Work 103

of tool-using involves the solution of 'diagnostic' problems, as those

faced by a physician. Z7 The matured craftsman of scientific inquiry
will be working on technical problems which are peculiar to him-
self and some colleagues, and which change fairly rapidly with the
development of his field; and so there is no substitute for his per-
sonal, largely tacit knowledge ofthe tools which have become nearly
continuous extensions of the sensory, motor, and intellectual
equipment within his body. The transmission of such knowledge
will then be largely through a close personal association of master
and pupil, and the explicit precepts of the refined methods of tool-
using have meaning only in the context of the solution of sophisti-
cated technical problems.
The informal and largely tacit precepts of method are not re-
stricted to the use of tools in the solution of the technical subsidiary
problems of a scientific inquiry. The scientist's craft also includes
the formulation of problems, the adoption of correct strategies for
the different stages of the evolution of a problem, and the inter-
pretation of general criteria of adequacy and value in particular
situations. These other tasks, which distinguish original scientific
work from the routine production of bits of information, have no
standardized, elementary versions to which simple, explicit precepts
can apply. Hence most of the body of methods governing this work
is completely tacit, learned entirely by imitation and experience,
perhaps without any awareness that something is being learned
rather than 'common sense' being applied. Since this sort of know-
ledge is so different in character from that embodied in the published
results, and is transmitted through a different channel, it is not
capable of the same universality of diffusion, nor of the same close-
ness of control of quality. It is therefore subject to particular weak-
nesses from which public scientific knowledge is protected. Yet it is
on this informal knowledge of the higher elements of the craft of
scientific inquiry that the success of the whole social endeavour
depends. We will later discuss the relations between these two sorts
of knowledge; for the present it is sufficient to establish the point
that in every one of its aspects, scientific inquiry is a craft activity,
depending on a body of knowledge which is informal and partly
tacit.
27 For an analysis of 'diagnostic' problems, and a demoDStntion that their structure
is similar to that ofgeneral scientific problems, lee M. Hamilton, Clifli&ias MUl Dedsitnu
(lDaugural Lecture, Leeds University Press, 1966).
104 The Achievement ofScientific Knowledge
Style
One of the reasons for the lack of awareness of the craft character
of scientific work is the form in which the results of that work reach
students and the lay public. There they are presented out of the
context of their creation and in a simplified or vulgarized version.
The fine structure of the argument by which the result was estab-
lished, and its context in earlier and alternative approaches, cannot
be conveyed except to a technically competent audience. It is just
in these fine points of detail that the style of a scientist is revealed. 28
The investigation of a scientific probl~m :s creative work, in which
personal choices as well as personal judgements are involved at
every stage up to the last. The scientist comes to the problem with a
unique set of interests, skills, and preferences. No two scientists
can set and solve exactly the same problem, except one which is
routine and sttaightforward. The materials of the problem are
malleable, and the conclusion which is achieved can vary in its
emphasis and form. Even when the problem is solved, its materials
are further developed by later work, in the context of new investiga-
tions. The apparently rigid and necessary form of achieved scien-
tific knowledge is arrived at only when the material comes into a
standardized version, for use as general purpose tools or informa-
tion.
Thus, even though the scientist is concerned with properties of
the external world, the work he produces will be characterized by
a certain style unique to himself. This style of scientific work is
analogous to that of artistic creation; historians of art have been
able to analyse works of art in terms of 'formal elements' of a com-
position, independent of the themes portayed. This sort of style has
been used as a foundation of the critical history of art, and through
this concept, links of influence (both within art and from outside)
have been established. 29 For the scientist, as well as for the artist,
21 For this section, I am embarrassed by a paucity of citations from the literature of
the history of science; and colleagues who are equally certain that style is an important
ISpect of scientific inquiry have been equally unsuccessful in finding substantiating
evidence. I am convinced that this silence is more a reflection ofthe traditions ofscholar-
ship in this field than of an absence of style from creative scientific work.
2' For a disc:wision on the use ofthe concept ofstyle in the history of political thought
in relation to its origins in the history of art, see K. Mannheim, 'Conservative Thought',
in ESMYS 011 SotioloD IIIUl SoeiIIl PsytalOD, ch. 2, 74-8. His concept of 'style' is one
characteristic of social groups rather than of individuals; but there is no contradiction
between his usage and mine. Closer to my own sense is the concept of style in teaching,
developed by M. Oakeshott, in 'Learning and Teaching', in R. S. Peters, Tile Contept
 Science as Craftsman's Work lOS

the personal style will be realized through choices within the range
of possibilities defined by the whole body of methods for his prob-
lem. There is no conflict between a highly individual style in the
investigation of problems, and the production of results which meet
the socially imposed criteria of adequacy for the field. Through
personal acquaintance with a scientist's style, a colleague can recog-
nize his work; and a later historian can describe the style, and even
use it for the explanation of a unique achievement. 30 Of course, the
development ofa personal style is a matter of degree; those scientific
workers who produce bits of information on order will not generally
need, or be permitted, a sttongly marked personal style for their
tasks.
Some important features of the social aspects of scientific activity
can be illuminated by the concept of style. Since the personal style
of a matured scientist is influenced by his earlier experience, we
can speak of the transmission of style from a master to his pupils.
In this way one can construct intellectual genealogies; and to under-
stand a man's work it may be relevant to know who was his teacher's

of EtlUCIII;tnI (Roudedge, London, 1967). '... in every "ability" there is an ingredient

of knowledge which cannot be resolved into information, and in some skills this may
be the greater part of the knowledge required for their practice. Moreover, "abilities"
do not exist in the abstract but in individual examples: ..• Not to detect a man's style
is to have missed three-quarters of the meaning of his actions; and not to have acquired
a style is to have shut oneselfofffrom the ability to convey any but the crudest meanings.'
(Q!1oted from T. H. B. Hollins, Anolher Looi III Tellther T,.tlini", (Inaugural Lecture,
Leeds University Press, 1969), pp. 19-20.) In the reporting of scientific results, the
style of investigation may not be allowed to dominate as in personal teaching; but for
the achievement of those results, without style there can be little originality.
30 A vivid account of personal style was given by Boltzmann, in connection with
Muwe11: 'A mathematician will recognise Cauchy, Gauss, Jacobi, Helmholtz, after
reading a few pages, just as musicians recognise, after the first few bars, Mozart,
Beethoven or Schubert. Perfect elegance of expression belongs to the French, though it
is occasionally combined with some weakness in the construction of the conclusions;
the greatest dramatic vigour to the English, and above all to MaxwelL Who does not
know his DytIII",;ca/ T1Jeory ofGtlSes? At first the Variations of the Velocities are de-
veloped majestically, then from one side enter the Equations ofState, from the other the
Equations of Motion in a Central Field; ever higher sweeps the chaos of Formulae;
suddenly are heard the four words: "put n= S". The evil spirit V (the relative velocity
of two molecules) vanishes and the dominating figure in the base is suddenly silent; that
which had seemed insuperable being overcome as by a magic stroke. There is no time
to say why this or why that substitution was made; who cannot sense this should lay
the book aside, for Maxwell is no writer of programme music who is obliged to set the
explanation over the score. R~t after result is given by the pliant formulae till,
as unexpected climax, comes the Heat Equilibrium of a heavy pi; the curtain then
drops.' (Q!1oted from R. A. MiIlibn, D. Roller, and E. C. Watson, Me,"",,"s, MoletUIM
Physics, Heal and Sound (Ginn & Co., 1937), p. 21ga.
106 The Achievement ofScientific Knowledge
teacher. 31 This is clearly an important element in the creation of
scientific 'schools'; and through the ttansmission of his personal
style a great scientist can have a direct influence on scientific work
even after he has departed and his own problems have become
obsolete. But this interpersonal action is also conditioned by a
style: that by which the scientist relates his personal scientific
endeavour to the communities of which he is a member. Whereas
the personal style of the scientific work will be largely conditioned
by the internal history of the field (perhaps as enriched by more or
less explicit philosophical concerns), this style of interpersonal and
social relations will be very directly influenced by the institutional
and social structure of the scientific community, and of the larger
society of which it is a part.
The particular character ofa scientific school will depend on these
two aspects of the style of its master. Since style is so largely tacit,
we do best to refrain from attempting an exhaustive list of its
components. But certain significant features of scientific schools can
be identified and explained on the basis of an informal understand-
ing of style. First, it is impossible for a personal style of scientific
work to be transmitted perfectly; for no two people, especially two
of different generations, can have experiences so similar as to pro-
vide bases for nearly identical styles. The difference between the
styles of master and pupil will be all the greater when the master
has done deeply original work, creating the objects and methods
of a field through his endeavours to attack some very fundamental
problem. For the pupil learns the developed forms ofthese materials,
and applies them to more straightforward problems; and in spite
of a complete personal allegiance, and an attempted copying of the
master's style, he cannot re-live his experience, and so cannot be a
replica of him. If the master has a certain style in his social relations,
he may fall into the pitfall of considering the brilliant pupil as a
true image of himself, fail to detect the lack of depth and subtlety
in his results, and so conttibute unwittingly to the vulgarization of
the products of his life's work. Other sorts of style in social relations
can lead to conflict, as when a master expects both brilliance and
obedience from his pupils in the execution of his grand programme.
31 H. T. Pledge, in Slime, sine, 1500 (H.M.S.Q., London, 1939), gives three charts
indicating master-pupil relatiODl in aperimental science from the sixteenth to the
nineteenth century. K. E. Ilothschuh, (;ese1lie1lt, tler Plysi%,;, (Springer-Verlag,
1953), gives several genealogical charts for this discipline.
 Science as Craftsman"s Work 10']
He will then find some disappointing his hopes.. and others betray-
ing his trust; and the continuation of his work through a school is
endangered.
Such situations of delusion, disappointment, and conflict do occur
in science, and of course they represent a less than perfect state of
affairs. But it would be a serious error to suppose that such errors in
perception and expectation by masters prevent the accomplishment
of great work, or even the formation of influential schools. For the
different strategies required at the different stages of the develop-
ment ofa field have different sorts ofstyle appropriate to them. The
special circumstances of pioneering work give rise to a style which
helps to explain a phenomenon frequently found in nascent discip-
lines. This is the presence of several schools, each loyal to the
exclusive message of its master, and all doing battle with each other
for several academic generations. Eventually, the passage of time
confutes all the warring sects, and later scientists freely use informa-
tion and insights from all sides. Then there is a general wonder that
people weren't sensible enough to do so at the beginning. But being
sensible about the variety of possible opinions and approaches is an
aspect of the style which is appropriate to a period of consolidation
of a field. The task of creating a discipline out of nothing is so
difficult and dangerous that only men with a special sort of personal
style will attempt it; the degree of dedication and courage which are
necessary for the work, frequendy carry with them a touch of
messianism. When such a man undertakes the heroic task of making
an objective science out of his deep insights into the truth, he can-
not afford to be broadminded about the claims of rival theories or
about criticisms of his own. It then requires only a strong per-
sonality and a favourable social and institutional situation for each
such pioneer to collect talented disciples, all of whose insights are
gleaned second-hand from the master, but whose personal dedica-
tion to his approach is, if anything, more rigid than his own. Each
such group then proceeds to erect a fortress of orthodoxy, and to
conduct battles until the field outside their enclaves develops to the
point that their conttoversies are merely boring. The direct descend-
ants within each original school, left with little but a rigid doctrine
and a scholastic style, then gradually wither away.32 Indeed, there
3a The mathematical sciences in the eighteenth century suffered from such 'scholasti-
ism', IS the British followers of Newton patriotically adhered to his clumsy notation
for the calculus, his one-sided views on the foundations of mechanics, and his theory of
the corpuscular nature of light.
loS The Achievement ofScientific Knowledge
occasionally occurs a paradoxical and dismaying situation where the
old master himself eventually loses his enthusiasm for the struggle
with the ancient enemy, and his disciples are forced to denounce
him as a renegade from the 'ism' bearing his name. Out of these
early conflicts within a nascent field there do appear results which
pass into the permanent record, appropriately standardized and
denatured; and it is left to historians to rediscover the drama and
tragedy which were an essential part of their creation.
 4
SCIENTIFIC INQUIRY: PROBLEM-
SOLVING ON ARTIFICIAL OBJECTS

Introduction
HAVING argued that scientific work is necessarily a craft activity,
depending on a personal knowledge of particular things, and on
subde judgements oftheir properties, I must now make an important
qualification. There are essential differences between the craft
work of scientific research and other sorts of human activities. For
the objects of scientific inquiry are of a very special sort: classes of
intellectually constructed things and events. Their difference from
the objects of handicraft production, or even of ordinary discourse
and action, gives scientific knowledge its special power, and makes
scientific inquiry a particularly complex and delicate social activity.
With this thesis I run counter to an influential tradition in the
philosophy ofscience, whose roots are in England rather than on the
Continent. It derives ultimately from Bacon, who believed that men
should, and could, turn to 'things themselves', and away from the
false theories used for describing them and explaining their actions.
A variant on this theme became popular in the nineteenth century,
through the writings of Huxley and Qifford: that science is nothing
other than 'organized common sense'.I With respect to the activity
of scientific inquiry, this belief in 'naturalness' is correct; in my
discussion of the different tasks and judgements involved in the craft
work producing data and information and using tools I did not find
it necessary to invoke any elaborate or formalized principles of
'method'. But in the extension of 'common-sense' from the work to
the objects of the inquiry, 1;he belief is seriously misleading. I shall
later discuss the very real and difficult problems of conttolling the
applications ofscientific knowledge; and these all revolve around the
• For example the classic work by W. K. cwrord, TM CtntIIIUJIJ Sense of1M uMl
SdnIe,s (1885; reprint, Dover, New York, 1955).
110 The Achievement ofScientific Knowledge
fact that this knowledge is esoteric, available only to those whose
long training in a field has made them expert.
Perhaps less crucial, but equally significant, this belief makes a
genuine history of science impossible. Until very recently at least,
the implicit assumption in the writings of the history of science
(especially at the level of schoolbooks and popularizations) was that
our sort of science is the obvious and natural way to study the world
of nature, merely the application of unfettered common sense.
Periods and cultures which failed to produce effective natural
science of our sort were condemned for an almost wilful neglect. By
the same assumption, great discoveries in science have been pres-
ented as the perception of the obvious, and the failure of a great
discoverer to get the thing quite right is the occasion for an apology
on his behalf. In this way, the history of scientific achievements
comes out as a succession of one-eyed men in the kingdom of the
blind. The question of why it required a great man to get even one
eye open, and still the question ofwhy he happened to come along at
just that time, are rendered incapable of being asked, let alone
answered. Such parodies of history are not merely of interest to
professional historians of science; since the published history of
science, especially at the popular level, is evidence for the self-
consciousness of science, it serves as a reminder of the necessity of
present discussion.
Intellectually Constructed Things and Events
I have already mentioned the classes of intellectually constructed
things and events which are the objects of scientific knowledge, in
my discussion of 'information'. Even in the production of data the
manipulations are with samples which are designed to serve as true
representatives of such classes, rather than with some 'real' objects
which can be known independently of an elaborated structure of
theory. Of course, these classes of things and events are intended to
relate, as closely as is possible, to the inaccessible reality of the ex-
ternal world. But they are different in character from that reality:
and from this difference derives the specialized nature of the craft
work of science, and the never-ending perfectibility of scientific
knowledge. 2
:I It is iDsufficiendy realized just how far removed are the conceptual objects, and the
problems, of scientific inquiry, from the world of ordinary experience. This absttaetion
can pass unnoticed because the teaching of science in the schools concentrates on the
accomplished knowledge concerning the natun1 world, with instances of where that
 Scientific /nlJuiry: Problem-Solving on Artificial Objects I II

An elementary example of such an artificial object is the concept

ofchemical 'substance'. The definition ofany particular substance is
highly complex, using properties derived from a variety of fields,
and always subject to gradual or rapid change. 3 Also, the 'substance'
is not a formalised description of a unique collection of material;
rather, it is a class of things, the members of the class being defined
by their possession ofcertain properties. Thus a chemical substance,
serving here as an example of an object of scientific knowledge, is a
class defined 'intensionally' by certain properties of its members. 4
knowledge has enabled effective conttol to be achieved. Also, since so much of the
environment perceived by people in advanced societies is man-made, those parts beyond
our oonttol or understanding are easily neglected. This abstraction was by DO means an
easy thing to achieve, and was not a feature of the 'science' of the early Greek philoso-
phers. See H. Gomperz, 'Problems and Methods of Early Greek Science', JMInIIIl of1M
History ofIilea, 4 (1943), 161..,6 (but not to be found in Contents or Index).
The great advances in physical science in the seventeenth century required an extreme
abstraction, and a putting to one side of outstanding problems, for its success. The
classic case of this is GaIileo's treatment of,motion' purely in terms ofa 'mobile' defined
by its changing position; the conttast with Francis Bacon's more traditional approach to
the study of motion, which proceeded through a taxonomy of motions, is striking. See
the NtJ1Nm Orgll"um, Book II, Aphorism 46; translation in Woris, iv, 214-32.
In the science of optics, the act Of abstraction was equally profound in its philosophi-
cal implications (which were quickly developed by Descartes), and in its OODSequences in
a one-sided development of the science over the following centuries. See V. Ronchi,
OptitS, tile Slimee ofVisitm, tr. E. Rosen (New York University Press, 1957). He observes
that the adoption of Kepler's hypothesis ofa telemetric triangle (the assumption that the
eye perceives the divergence of rays from a point source) enabled the physiologico-
psychological aspects of vision to be ignored thenceforth (p. 50). In the book he shows
'how productive of error it was to have forgotten that optics is the science of vision and
to have developed it blindly (for that is the appropriate word) as the optics or images or,
I reiterate, as the optics of the centtes of the waves' (p. 206). For an extended historical
study of optics before and after the onset of this blindness, see V~ Ronchi, Histoir, tlt •
Lumib, (S.E.V.P.E.N., Paris, 1956). A summary of his thesis on the history ofoptics is
'Complexities, Advances and Misconceptions in the Development of the Science of
Vision: What is being Discovered ?', in Stimliji& C/umte, ed. A. C. Crombie (Heinemann,
London, 1963), pp. 542-61.
3 The element 'Oxygen' has undergone such changes since its first discovery, or,
better, invention. For an acoount of the changing conceptions of this substance and its

C"""y lind llIe First P.,

properties, see E. Farber, Oznnl MIl OzitUt;ort-Tlleones IIfIIl Teelmitws i" tM 19t'
of 1M 20tIJ (Washington Academy of Sciences, Washington,
D.C., 196'7); reviewed in Isis, 59 (1968), 454-5. The concept of 'acid' with which 'oxide'
was identified by Lavoisier has also evolved since the identification was shown to be
false. For a brief sketch, see J. W. Mellor, A Compr,1Iensiw T"Mise 011 IfUWgait ••
Tlleoretittsl Chemistry, (Longmans, Green, London and New York, 1922), 385-6.
4 For a penettating philosophical analysis ofthe concept of element, see F. A. Paneth,
'The Epistemological Status ofthe Chemical· Concept of Element', BritisA JMInIIIl for 1M
PlJilosoplJy of Slimee, 13 (1g62-3), 1-14 and 144-60; especially sectiODl 5 and 6, pp.
149-60. (This is a translation from the German; the original was published in 1931.) A
similar theme, with a wealth of historical examples, is discussed by R. Hooybas, 'The
Concepts of "Individual" and ~'Species" in Chemistry', C"""unu, 5 (1958), 307-22.
112 The Achievement ofScientific Knowledge
To determine the 'extension' ofthe class, seeing whether a particular
object, given independently, belongs to it, we must decide whether
this object is a 'sample'. For this, we test whether this real collection
of material satisfies the defining properties of a member of the class
to an acceptable degree, and which moreover contains specified
'impurities' (which would render it a 'non-sample') to no more than
a tolerable degree. We notice that although the 'sample' cannot be a
perfect realization of a member of the class which constitutes the
defined 'substance', it is still in itself a highly artificial creation; few
'chemicals' can be found in the world of nature, outside the labora-
tory.
The same example serves to show how scientific knowledge is
concerned with such classes of things and events, and not with
particular things and events that happen to be observed. Demon-
strations in school science courses are explicitly designed to show the
properties of the general classes, through the behaviour of selected
representatives. The merriment that ensues when a demonstration
'goes wrong' proves this; the teacher has failed to show the pupils
what 'really' happens to the samples representing the general class,
and has instead produced an effect involving extraneous and un-
wanted other things and events. It might be thought that the more
descriptive sciences are closer to the real things of nature, and are
not so completely imprisoned in the world of their own concepts.
But the history of the centuries-long search for a 'natural' system of
classification (and hence description) of living organisms shows that
here too the objects of scientific knowledge are intellectually con-
structed classes rather than 'things themselves's
Chemical 'substances' and biological 'species' are less artificial
than many objects of scientific knowledge, for their samples are
things which have many properties accessible to fairly direct in-
spection. Many conceptual classes have samples whose properties
S An example of the artific:iality ofa basic concept in a social science is that of 'price',
used freely by economists as a simple quantitative variable in many contexts. O. Morgen-
stern, 0. 1M A,ewflty of E,tmOtIIi, Ol1servtllioru (2nd ed., Princeton University Press,
1963) shows how ndical an abstraction this is from the complex realities of the terms of
trade (pp. 181-7). Hard-headed businessmen fare no better; 'profit' is revealed on
analysis to'be a mOlt elusive concept, whose measure depends more on book-keeping and
accounting mnventions than on any particular aspect ofthe actual opentions ofthe firm
(pp. ,0-8,). The more elaborate measures, compounded out of aggregates and estimates
of these basic ones, such IS 'national income' and 'gross national product' and 'growth'
have a contact with reality which is far less firm than the precision of their estimates
would suggest.
 Scientific /nlJuiry!· Problem-Solving on Artificial Objects 113

or even whose character as a 'thing' or 'event', are incapable of easy

ttanslation into ordinary experience; 'field' is a good example ofthis.
In such cases, as in that of such basic entities as 'mass' and 'energy',
it is beside my present purpose to determine whether the scientific
concept is best considered as a class of 'things' or of 'events', where
these terms are taken over from ordinary speech. For to the extent
that their ontological status becomes obscure their character as in-
tellectual consttucts becomes apparent; and in the same measure,
any particular manual experience involving them becomes sophisti-
cated and theory-laden.
It might be objected that I have not yet shown any distinctive
feature of scientific knowledge, for even our ordinary speech and
thinking are done with names, which are idefitifying tags for general
classes ofthings and events. Moreover, these names, and their classes,
are far from natural; not merely absttact concepts, but the names
and classification of everyday things are the result of cultural pro-
cesses, varying widely from one milieu to another, and constandy
changing over time. In this sense it is ttue that all our knowledge is
'artificial', and that we cannot conceive things for which our own
culture provides no language. But the objects ofscientific knowledge
are even more artificial than this; to indicate the difference I have
used the term 'intellectually consttucted'. We can see the difference
when we consider the border areas where terms from the yemacular
are taken over into disciplined discourse. This occurs when the
looseness of ordinary speech, and common sense thinking, is in-
adequate to the needs ofthe situation; public administration and the
law are familiar examples.
In -these fields, the intricacy of the relations into which people get,
must be grappled with; and procedures are developed so that com-
plex practical problems can be solved in consistent fashion. Much
ofthe work of developing administrative or legal systems can be seen
as defining classes of things and events, and setting the decisions to
be taken in situations involving members of those classes. The
routine work of decision-making then consists largely in determining
the classes to which the real people, and their particular tangles,
should be assigned. These intellectually consttueted classes derive
ultimately from common sense experience; but in their detailed
elaboration, the need for systematic coherence weighs at least as
heavily as the retention of the original link. In any sane system of
this sort, there is a place for common sense ideas to be injected
114 The Achievement ofScientific Knowledge
directly into the decision-making process; but this is righdy kept as
a last resort, or else the whole system would collapse, and chaos
would supervene.
The necessary artificiality of developed legal or administrative
system is analogous to that of a developed science. The world of
nature is far too complex to be comprehended in terms of the con-
cepts we build up in our ordinary experience. Nor can we interact
directly with the natural world, without the mediation of our
organizing and simplifying conceptual structures. Of course, the
screen between us and the natural world, which is necessary if we
are to achieve any effective knowledge of it, can also serve to shield
us from an awareness of our ignorance. Just as a legal or ad-
ministrative system can drift out of all contact with the social
realities it is intended to service, a scientific discipline can become
totally self-contained and sterile. More seriously, its practitioners
can come to take their system of artificial objects as the basic
reality, ignoring its differences from that external world which it
was originally intended to represent. 6 A fuller discussion of these
social problems must wait until we have seen in more detail how
the artificial nature of its objects influences the craft work of scien-
tific inquiry, and gives it its peculiar sttength and delicacy.

The Complexity of'Discovery'

When we accept the view of science as craft work operating on
intellectually constructed objects, some traditional views on the pro-
cess of the achievement of scientific knowledge are immediately
excluded. Because of their importance, I shall discuss them briefly
6 The tendency of established scientific disciplines to create a world of their own and
then live totally within it, is now becoming recognized as an obstacle to the solution of
pnctic:al problems involving science and technology. Thus, 'I must also stress the
incompetence of the established disciplines to tackle society's real problems. What we
mean by a discipline is an agreed, tested body of method-usually analytical-that we
bring to bear on problems of ow 0"", ,hoosi",. The essence of our thinking is that we
cannot tackle problems that don't fit the competence ofour own discipline. It's true that
we constandy try to enlarge that competence. Confronted with a new problem, we spare
no dort to improve our methods. But if we don't succeed, we don't tackle the problem,
and we tend to condemn mlleagues who try.' J. Kenneth Hare, quoted in John S. Stein-
hart and Stacie Cberniak, Tile Universities.ruJ EmJirtmlllnltll1 Qulity-C01IIItIitmmt to
ProbleM Fonuetl Etlll&";(JfI-A Report to tile Pr,siIlmt's E"virtmment"l QUIIl;t;, Co""eil
(United States Government Printing Office, Washington, 1969), p. 6. For a review of this
significant document, in which the philosophy of science developed here is implicidy
presupposed, see L. J. Carter, 'Environmental Studies: OST Report Urges Better
Effort', S,imee, 166 (1969), 85 1•
 Scientific /nlJuiry: Problem Solving on Artificial Objects lIS

before going on to an alternative approach, in which the special

character of the craft of scientific inquiry is described.
The ttadition of the 'naturalness' of science involves, as we have
seen, the assumption of the ttiviality of scientific discovery. In
principle, all one needs to do is to go out with an open mind and
clear eyes, and one will find facts and discover true laws of nature.
Such faith doubtless had its social and ideological functions at some
times in the past, but it is a travesty of the work of achieving scien-
tific knowledge. It is possible for a master-craftsman (especially in
descriptive science) to make observations and inferences with a
speed, and apparent ease, that astonish his less gifted and less ex-
perienced colleagues. But such discoveries come from the opposite
ofan open mind and clear eyes; rather, they come from a mind which
is constandy, if quietly, at work on a multitude of problems; and
from eyes which are trained to perceive phenomena which, when
interpreted in terms of the artificial objects of his discipline, are
significant for the solution of those problems. Also, such a master, if
he does work worthy of his talents, will be engaged on really big
problems, and the deeper results he achieves will be reliably estab-
lished only by years of arduous labour.
Any serious study ofthe history ofscience destroys the assumption
of the simplicity of a scientific discovery. What we find instead is a
brief phase when the objects of inquiry are in rapid change, and
are correspondingly complex, ill-defined, and plastic; until one
person achieves a solution ofa problem which seems to determine the
objects decisively, but which leaves further work to be done until
the 'discovery' is complete. Perhaps the best example of this process
is the classic 'discovery' of oxygen by Lavoisier. At around the same
time, Scheele and Priesdey isolated a substance which they recog-
nized to be of importance; they named it 'fire air' and 'eminendy
respirable air' respectively. It was Lavoisier who realized that this
substance was the master-key to a new system ofchemical theory and
nomenclature; he named it oxygene, and the glory of the discovery is
his. Unfortunately, the intellectually constructed class so named by
Lavoisier is not the simple ancestor of our present element Oxygen.
For in his system of chemistry, it was a 'principle' responsible for
two classes of phenomena which he hoped thereby to coalesce: the
burning of substances in air, and the formation of acids. And it was
as 'acid-maker' rather than 'fire-maker' that he named his principle
(as one sees clearly in the German translation of Sauerstoft); and so
116 TIJe Achievement ofScientific Kno1lJletlge
Lavoisier's error in the creation of this object is preserved in the
name by which his 'discovery' is enshrined. 7 Historians ofchemistry
have not quite come to grips with the problems of judgement in-
dicated by this classic error; those who opposed the new French
nomenclature on the grounds of its being theory-laden are still
dismissed as 'reactionaries' in authoritative histories. 8
Indeed, one could argue that no discovery of any sort can be made
of 'things', but must always involve an incompletely specified intel-
lectually constructed object. The name involved in the classic
question 'Who discovered America?' indicates that this may be a
paradigm example for such a thesis.

Scientific l1UJUiry tiS a Ttlsk

We can now attempt to combine the materials of the two themes
we have developed so far, and discuss scientific inquiry as a special
sort of craft work, which operates on intellectually constructed
objects. Out of this will come an anatomy of a 'scientific problem',
and the possibility of giving a useful definition of this concept to
serve for our subsequent discussion.
If we are to proceed by analogy with craft work in general, we
need an anatomy of the tasks which are accomplished in that sort of
activity. This is available in the writings of Aristode. His scheme is
most popularly known in terms of four 'causes': final, formal,
efficient, and material. But it is clear from his own examples that
these are the constituents of the task, which includes the work and
its objects. The 'material' cause is the physical substance which is
worked on; the 'efficient' cause is either the agent or his activity of
shaping it; the 'formal' cause is the shape which is realized in the
work; and the 'final' cause is either the purpose of the activity (the
creation ofa specified object) or the functions of the object itself. For
7 See T. Kuhn, 'The Historical Structure of Scientific Discovery', Stimet, 134 (I June
1962), 760-4, for an illuminating discussion of this example. Historians of chemistry
have generally avoided emphasizing this error of Lavoisier, and have concenttated on
his studies of combustion. But if Lavoisier had considered combustion rather than
acidification as the essential property of this element, he would doubdess have called it
PFOIN nthcr than oZYIN.
• Thus, M. P. Crosland, Historie.l Sttulies in the unguage ofChnnistry (Heinemann,
London, 1962) mentions the 'reactionary views ofthe Editor oCthe Observal;om' (p.IS9);
this was de Ia Metherie, who objected to the new nomenclature on the (correct) grounds
of its being misleading (not all acids being oxides) and involving 'hard and barbarous
terms'; and he published letters showing (correctly) that the new nomenclature was
theory-laden (in rejecting the phlogiston theory of combustion).
 Scientific /nlJuiry: Problem-Solving on Artificial Objects 117

Aristotle, there was an intellectual and social hierarchy in these

constituents, downwards from the 'final' cause; for this is the one
whose setting is a policy decision at the highest level, and it then
determines the less independent and less intellectual work of making
plans and then shaping bits of passive matter. 9 Aristode used the
principle 'art imitates nature' to adapt this scheme to 'natural'
productions; that is, processes which come to completion by their
own internal plan, without the need for human intervention. With
this, he could make his fundamental criticism of the earlier 'atomist'
philosophers: that in concentrating on the efficient causes of the
production of natural things, they had never inquired into what is a
'thing'.IO On the other hand, Aristotle made a complete separation
between the production ofthings and the achievement ofknowledge.
For him, knowledge was concerned with 'what could not be other-
wise', and hence was eternal and unchanging; it had nothing in
common with the production of things, which is in the realm of the
variable and the contingent. 1 1 He may have been influenced in this
by his training under Plato, and also by his briefhistorical experience
of scientific knowledge. It is only in the twentieth century that we
have become familiar with the historical evolution of scientific
knowledge, in which great scientific achievements which seemed in
their own time to give true and necessary results are inevitably
modified and superseded.
We can adapt Aristotle's scheme to our own purposes by identi-
fying the differences between the objects of the two sorts of craft
work: making material things, and scientific inquiry. In our case,
the task is not accomplished by the production of an object which
has a certain function, and an appropriate shaJ}C imposed in certain
sorts of matter. Instead, it is a statement of a particular sort. The
apparent form of the statement can vary widely: it can appear as a
report ofan experiment or observation, a description and analysis of
a complex situation, or as a statement of a fact of or a law, or as an
hypothesis, theory or model. This list could be extended; but for
our present purposes, the differences between these types of state-
ments are less important than what they have in common. First,
their 'materials', the objects to which they refer, are always the
9 For a discussion of the basic role of the 'craft analogy' in Aristotle's thought, and
its roots in earlier Greek philosophy, see F. So1msen, 'Nature as Cnftsman in Greek
Thought',:/fJunuI1 oftbe History of IMtIS, 24 (1963), 473-96, especially pp. 48str.
10 See the D, P.tilnu A";",.lium, Book I, ch. I, 642& 2S.
II See theElm,,, Nie01fllMbe", Book VI, m. 2, 1139b 18-36.
118 TIJe Achievement ofScientific Kno1lJletlge
intellectually constructed classes of things and events. Particular
instances are reported only as representative samples of those
classes. Second, the 'form' ofa completed piece ofwork is not a bald
assertion that such-and-such is the case; rather, it is an argument, in
which evidence is cited and from which a conclusion is drawn. In
the discussion of the evidence, there will be some description of the
relevant part, 'efficient cause', by which the evidence was produced.
Under this will be included mainly the tools and techniques used;
there may (depending on the conventional literary style for research
reports in the field) also be mention of the difficulties and pitfalls
encountered and overcome. Other aspects ofthe efficient cause which
are usually omitted in research reports are a history of the investi-
gator's work, and the more subtle methodical judgements which
guided and conttolled his work. And last, the statement which is the
goal of the work as a whole, the establishment of new properties of
the objects of the investigation, emerges as a conclusion from the
argument.
Thus our first sketch of an analogy with the Aristotelian scheme
yields as the 'material' constituent the classes of intellectually con-
sttucted things and events which are the objects of investigation;
as the 'efficient' constituent the agent, with his work, tools, and
methods; as the 'formal' constituent the argument; and as the 'final
cause' the conclusion ofthe argument, with its new statements ofthe
properties ofthe objects ofinquiry. The data and information which
provide the sole contact with the external world are not the con-
clusion and are quite distinct from it. They are nothing more than
the foundation for the evidence, which concerns the objects of
inquiry, and which is embedded in the argument about them. These
different constituents of the completed work are interrelated, and
they condition each other. The character of the 'material' for a
given problem delimits the sorts of operations and tools which can be
applied to it for the eventual production of evidence, and corres-
pondingly delimits the sorts of arguments which can be consttucted
around that evidence. Conversely, the setting ofa particular problem
within a field will determine (within the above limits) the sort of
data which will be brought into being, and the particular sorts of
tools and arguments which will be brought to bear.
The artificiality of the objects ofscientific inquiry is the key to the
deep differences between this sort of craft work and that which
Aristotle took as his paradigm; hence we can best start our closer
 Scientific Inquiry: Problem Solving on Artificial Objetts 119

analysis of this scheme with an examination of the 'material' com-

ponent of a scientific problem. It is the establishment of new
properties of these intellectually constructed classes that is the goal
of the task. In this, the objects themselves are altered, for they exist
only as classes defined by their properties. We should recall that not
all their properties are exhaustively defined by the formal statements
presented in the public record; each scientist must have a crafts-
man's intuitive, personal and partly tacit knowledge of his intellec-
tual objects and of their physical samples, if he is to work creatively
or even competently with them. But there must be a large common
core of practical knowledge of the objects (again, not all of it in the
explicit record) if there is to be an effective social endeavour of their
study.
The study of these objects, which are intended to correspond to
an inaccessible external reality, cannot progress very far without
some interaction with the ultimate sources of our knowledge in that
reality. But the fashion in which this interaction yields new properties
of these objects is highly indirect and in principle inconclusive. We
have already seen that the production of data, and its refinement
into information, are craft operations which are governed by the
problem at hand, and which produces less than conclusive infer-
ences from the behaviour of particular samples in unnatural con-
ditions. The connection between such a foundation in particular
experiences, and the establishment of the properties of intellectu-
ally constructed classes of things and events, is in principle tenuous.
And in practice, it has been forged only after long periods ofdevelop-
ment, extending over generations or centuries for the different fields
of scientific inquiry.

The Argument
This link between the 'material' component and the 'final cause'
of a problem is formed by the argument. Through it the experience
is made relevant to the properties of the objects of inquiry. The
argument will be a lattice-like structure of assertions about the
objects of the work, connected by inferences which are accepted for
the particular linkage functions they perform. Each assertion must
be based, either directly or indirectly through these inference-links,
on a statement ofexperience or on an explicit postulate. We shall see
later (in connection with the 'adequacy' of solutions of problems)
that the argument must have a structure that is both complex and
120 The Achievement ofScimtiji& K"ollJledge
subtle; for there is no formally valid pattern of argument that can
establish properties of general classes from reports of particular
experiences. The argument may be partly mathematical and deduc-
tive, but (outside purely theoretical fields) it must also include
inductive, confirmatory, probabilistic, or analogical inferences,
which are never capable of carrying certainty from premiss to
conclusion. l :& Because of this lack of demonstrative certainty, the
argument will include subsidiary arguments on the strength of
particular inferences, especially for those that have a crucial position
in the structure.
First, they cannot yield true and certain conclusions about the
objects of inquiry, and still less about the external reality which they
are intended to represent. This limitation is revealed by the sub-
sequent history of any solved problem, even those which seemed (in
their popularized versions at least) to establish irrefutable properties
of natural things; I shall discuss this at greater length in connection
with the evolution of scientific knowledge. Second, the necessary
complexity and subdety of a scientific argument makes impossible
any trivial testing ofits adequacy, as is possible through the examina-
tion of schematic structure in the case of simple syllogisms.
Rather, the testing is done by an application of the criteria of
adequacy accepted for the field. This itself requires judgements;
and I shall discuss it more thoroughly in connection with
'methods'.
At first glance, an argument presented in a research report may
appear to have a pyramidal structure, with a number of reports of
experience, citations of known information, and perhaps postulates,
all leading to a conclusion which consists of one or a few assertions.
But this reflects the conventional style for such reports, rather than
a real completeness of the argument. At a variety of points, the
structure of such an argument could be developed to yield further
conclusions; and one of the paths to the investigation of descendant
problems involves doing just this.

12 The complete set of such inferences is present, if implicidy, in any scientific paper
where experimental results are obtained and also explained. Inductive inferences are
involved in the taking ofthe sample studied as a ttue representative of its class; probabili-
stic inferences are involved in any assessment of the 'goodness of fit' of data; deductive
inferences are the basic links in a verbal or mathematical argument; analogical in-
ferences are used for relating the concepts understood by the mathematical symbols to
the objects ofinquiry involved in the experiment; and confirmatory inferences are made
when the evidence is cited in support of the conclusions of a theoretical argument.
 Scientific Inquiry: Problem-Solving 0" Artificial Objects 121

Evidence
Only in the most purely 'descriptive' of fields can the information
have the appearance of being the direct and sole foundation of those
statements about the objects of inquiry which comprise the con-
clusion. Generally, the information appears as embedded in the
argument, providing a part of the basis for the conclusion. Hence
this information in the argument is not a statement of the properties
of the objects described in the conclusion; rather it is evidence
brought into the argument which as a whole establishes these
properties. The information which is selected from the available
stock for use as evidence must be subjected to further testing, not-
withstanding the fact that it was originally produced with this
function in mind. For each use of information as evidence is special
and demanding. Since no piece of evidence based on reports of
particular things and events can entail a positive assertion about a
general class, the weaker inferences which are made must be scru-
tinized in the light of the work they do.
The evidence must be of sufficient strength to support the load of
argument placed on it in its particular location. Moreover, the
evidence must be shown to fit the statement being supported, in
the correspondence of the objects of its reports with those of the
statement. A failure to make a proper assessment of the strength
and the fit of evidence before including it in the argument can lead
to pitfalls in the drawing of the conclusion. For these assessments,
the scientist again makes use of the criteria of adequacy accepted for
the field; and the interpretation and application of these criteria to
each particular case involves making judgements, using the craft
skill of the scientist.
The special category of 'evidence' is most easily recognized in
fields where problems involve both complex arguments and large
masses of information; and where the information itself does not
bear obvious credentials of its reliability and relevance. An extreme
example of this situation is in the law, where there is a highly
developed 'law of evidence' for the presentation and testing of in-
formation offered as evidence in court cases. In the disciplines dealing
with human history, the pitfalls which beset the inferences made
from information in its use as evidence are a recognized hazard, and
so any crucial piece of evidence must be carefully scrutinized for its
strength ilnd its fit. In most work in the natural sciences, one usually
has either a large mass of information with a relatively simple
122 The Achievement ofScientific KtJOlIJledge
argument, or a complex argument needing evidence at only a few
points. Hence neither 'descriptive' nor 'theoretical' natural sciences
generally require highly developed skills in testing evidence beyond
the tests for reliability and relevance already involved in producing
information. As in any other sort of'work, when there are fewer pit-
falls encountered and fewer special skills required the achievement
ofreliable results becomes easier; and hence more effective work can
be done by the same level of talent. Thus one reason for the greater
power of the sciences of nature, compared to most of the sciences of
man, lies in the simpler character of routine work, and the corre-
spondingly greater competence of mediocrity. 13
But even in such well-established fields pitfalls can be encountered
in the interpretation and use of evidence. For example, the con-
temporary fashion for using mathematical materials at every possible
point of an argument induces a tendency to accept statistical in-
formation as facts, rather than as evidence, in a wide variety offields.
This can lead to negligence even in testing the reliability of such
information in relation to the populations whose properties it
describes. The pitfalls in the way of even the humble tasks of
statistical data-collection are well-known to those who make their
living by the quality of the inferences based on such materials (as
commercial market surveys) but not always so well-known to others.
Moreover, the relevance of the classes established for the collection
and processing of data to those of the argument ofthe problem is by
no means automatically assured. Even when the problem is a purely
'empirical' one of determining a statistical property of a population,
13 The classic paper on the determination of the values of physical constants, R. T.
Birge, 'Probable Values of the General Physical Constants', Revietl1s ofMotlem PAys;,s,
1 (1929), 1-73, shows how subtle is the evidence, and how complex are the arguments, in
really fundamental physical research. In his introduction (pp. 1-7) he lists the common
pitfalls in the conversion of experimental data to what I call 'information': mnfusion in
the sense of 'experimental error'; the fitting of curves or lines to graphed points by
estimation nther than by calculations; and the use of incorrect values of the 'auxiliary
constants'. The argument will involve a discussion of the known sources of inaccuracy
in the data and information, as well as mmplex calculations for deriving the final results.
He stresses that the experimenter (as well as the reviewer) must exercise his judgement
at every stage. Finally, he observes that some painstaking experimental work, which had
produced accepted results, was in fact vitiated by one or another error: and, in some
cases, it was even impossible to 'save' the experimental data because the loss of the
records of its computation (including special techniques and values of auxiliary mn-
constants) prevented a ~working of it. In D. Lerner, Evillmte llrulI"fermee (The Free
Press, Glencoe, 111., 1959), there is an essay on 'Evidence and Inference in Nuclear
Research' by Martin Deutsch (pp. g6-106) which shows how indirect and mnditioned
both by theory and by expectation is the data of hig~ergy physics.
 Scimtiji& /tllJuiry: ProlJlem-Solving on Artiji&;a/ Objects 123

there must be an argument establishing the quality of the

information; and if any inferences whatever are drawn, there
must be supplementary arguments for establishing the strength
and fit of the statistical evidence for its function in the main
argument. 14
The pitfalls to be encountered in connection with evidence are
most noticeable when existing information is taken over for use in a
problem. Whether this material is immediately capable of an
adequate strength and fit for its function in the argument will
depend on the mode of its original production. For when it is cited,
the argument by which it was derived is implicidy included in the
argument of the new problem, as a subsidiary argument for its
adequacy. Also its objects are implicidy assumed to be those of the
new problem. If the original objects were significandy different
from the new ones, or if the original argument was of a pattern
which is not adequate or is inappropriate for the new problem, then
the information itself is inadequate for its new functions as evidence
and the new argument fails. In a general way, these considerations
lead to a precept of method, of using only that existing information
which is at least as solidly based on experience as the new informa-
tion of the problem at hand. As an extreme case, one does not use a
hypothetical conjecture as crucial evidence in a largely inductive
argument. But the pitfalls do not lie exclusively in shifts in this
direction; for the objects ofinquiry in problems closer to experience
will frequendy be different from those of the same names in more
theoretical investigations; and so even 'empirical evidence' may be
irrelevant and misleading to the inquiry at hand.
The neglect of the category of evidence clearly derives from the
traditional concerns of the philosophy of science, and its paradigm
examples of scientific achievement naturally obscure this aspect of
scientific work. The classic inclined-plane experiment of Galileo
has served for generations to exhibit science as the sttaightforward
and simple production of data which immediately becomes fact. In
recent years, historical scholarship has shown how deeply theory-
laden was this particular experiment, in its concentration on very
14 On the abuse of national statistics for economic policy-making, see A. Coddington,
'Are Statistics Vital?', TAe Listener, 82 (1969), 822-3- He deals DOt only with the un-
certainty in the sampled data and scaling facton whereby the gross statistics are calculated,
but also with the more subtle questions of the effect ofthe systems ofconventions on the
production of the data itself; in this last respect his analysis goes beyond that of O.
Morgenstern.
124 The Achievemmt ofS'ientiji& KtJOlIJledge
artificial and novel objects. IS Moreover, there could not have been
chosen a more misleading example ofGalileo's approach to the use of
experience. An attentive reading of Galileo's major works will show
that he was fully aware of the problems of evidence, and of the
qualities of strength and of fit of evidence deriving from various
sorts of experience for various sorts of argument. Q!tantitative
experiments actually performed were very much the exception for
him; his evidence derived from a continuum of sorts of experience,
ranging from surprising phenomena and craft experience, through to
abstract thought experiments. 16 His conclusions show a corre-
spondingvariety, from the initial framing of problems to be solved,
through the tentative or speculative solution of problems, to rigid
mathematical demonstrations on expIicidy artificial objects.
When we leave the world of routine scientific work by competent
craftsmen, the problems of evidence change radically. In any genu-
inely novel work, or even in work of crucial importance for a big
problem, the accepted criteria of adequacy may not extend to cover
all the inferences made from the evidence in a particular argument.
For the novelty of the problem will generally entail a corresponding
novelty, and uniqueness, in evidence and its relation to experience
and to the argument. Also, since such work is frequendy done at the
15 The pioneering work on this interpretation of GaIiIeo was done by A. Koyre,
£lIIIks GlIlilletlMs (Hermann, Paris, 1939). Koyre even argued that Galileo ~d not
have achieved the precision he claimed. But the doubts of Galileo's French mntem-
ponries whichKoyre cites (0, 73) are not on the distanc:e-time law, but on other assertions
of Galileo, published in the Dilliogue of 1632; see R. Lenoble Mersmne, 011 U N.UlltI&e
tl. Mle"';sme (Vrin, Paris, 1943), pp. 465-6. For a brief study of this as part of Galileo's
work, see The Mathematical Association,.A SeetmtlReport tnltM TeMkin, ofMe,"",",s in
, . SeluJols (Bell, London, 1965), ch. 6, 'Problems and Methods in Mechanics---u
Historical Sketch'.
16 In the TJI10 NnIJ Stinlees we find surprising phenomena in industrial practice
mentioned at the very beginning as a fertile source of problems (&lizitme NflzitnJale,
P.49). Among these is the 'size effect', the rule that a large ship under mnstruction would
faD apart under its own weight unless supported, while a small one of the same design
is stable. Craft experience is the basis for the observation that a rope can be made to grip
on the smooth drum ofa windlass (p. 58), that a rope breaks by the tearing, not slipping,
of its strands (p. 56), and that ropes can lift heavy weights in being shortened through
wetting by mist (p. 6,). Galileo's experiments include the classic quantitative ones;
qualitative experiments such as the adhesion of two polished plane surfaces (p. 59); and
experiments that mold not succeed, as that of a piston loaded with a weight until it
descended against the 'force of the vacuum' (p.62). Thought-experiments abound in the
Tllirtl DflY; one shows that an impercepbole height of fall produces an imperceptible
impact (p. 2(0). Finally, Galileo solves some problems by 'pure reason alone'; thus he
argues against an instantaneous assumption of a determinate speed by a body beginning
to fall, on a principle of'suflicient reason': DO speed is more likely than any other (p. 2(0).
 Scientific Inquiry: Problem-Solving on Artificial Objects 125

limit (if not beyond) of the capabilities of the existing tools for
producing reliable information, the assessment of the strength and
fit of the evidence can become very subde indeed. l ? Then there can
arise disputes about the adequacy of the solution to the problem,
which cannot be resolved either by a scrutiny of the data and in-
formation, nor by an appeal to accepted criteria of adequacy. Thus
at such points, this aspect of the 'objectivity' ofscientific knowle4ge,
which is really a result of a successful social ttadition of producing
and testing the materials of that knowledge, breaks down. In the
long run, to be sure, further work will decide the issue; but the
decision on whether to engage in such further work, which partly
depends on the assessment of the adequacy of the controversial
piece, must be taken now. Thus, at such infrequent but critical
junctures in the advancement of science, the assessment of the
evidence adduced in an argument becomes a crucial judgement, in
which the individuals are thrown back on their own personal re-
sources. They are forced to put themselves at risk in making the
judgement, and they lack the safe channels of an accepted ttadition
to steer them towards the correct answer.

The Conclusion
At the end of an argument, comes the conclusion; and this is the
completion of the cycle of a research project, which is the first step
towards the achievement of scientific knowledge. The artificiality
of this product of craft work should by now be obvious. The
conclusion is not concerned with 'things themselves', but with those
intellectually constructed classes which can serve as the objects of
an argument. The contact with the external world is always un-
natural and indirect; and the reports of that contact do no more
than serve as the basis for evidence which is embedded in an argu-
ment whose pattern cannot be formally valid. Although the state-
ments in the conclusion refer to explicit objects, and are capable of
being understood by any competent practitioner, and although
every stage in the argument, including the reports of experience, can
17 Pasteur's great discovery of left and right-handed crystalline forms of tartrates was
exceptionally fortunate in its circumstances; Pasteur could convince the influential Biot
of its validity by an experiment performed in his presence. Otherwise, he might have bad
difficulties, for it is only under extremely carefu1Iy controlled conditions that the chemical
reactions needed for producing the effect can go through properly; and crystals of the
sort observed by Pasteur are very rare. See A. Ihde, Tile Dewlo",."., ofModml C1mlti1lr~
(Harper & Row, 1964), pp. 322-3.
126 The Achievement ofScimti./i& K,,01lJledge
be reproduced or tested, the work as a whole has been conditioned
by personal judgements, depending ultimately on a private, craft
knowledge.
Whether such an account ofthe completion of the investigation of
a problem seems a reasonable one will depend on the experience of
the reader. For those who imagine science as the accumulation of
hard facts and indubitable truths, a 'conclusion' as described here
will appear to be a miserably weak result, a parody of the achieve-
ment of scientific research. Indeed, it is hard to imagine how the
magnificent edifice of our established scientific knowledge could
be composed of such feeble elements. We will later see that in fact
it is not so constituted; for the conclusion of the argument, along
with the other components ofa research report, enters a new cycle of
development, in which it is further tested, and also transformed, so
that its weaknesses are exposed and if possible corrected.
But for those with either an experience of scientific work in any
lively field, or some knowledge of the history of science, this de-
scription should seem a fair one. Indeed, through an understanding
of the inherent limitations of the conclusion of a problem, we can
appreciate the inevitability of error in scientific work, at least in the
productions ofall but the men ofthe greatest genius. For a conclusion
can be no better than the evidence on which it is based, and the
objects in whose terms it is framed. The objects of enquiry delimit
the sorts of problems that can be set; and to the extent that these are
capable of enrichment and deepening, then will later problems be
correspondingly improved, and their solutions more powerful.
Similarly, the evidence depends on these, and on the tools available
at any time; and these too improve. When there is a debate over a
scientific result, neither side can draw a blank cheque on what might
be produced in its favour were better tools available; the issue must
be decided on the evidence that is there for scrutiny. A new set of
tools, providing new and better evidence, may well tip the balance,
so that while it was reasonable and correct to support one solution
under the old conditions, the new ones require a change of position.
Such changes are not easily made by a scientist in mid-career, of
course: but that is another problem, of a practical sort.
From this argument, we can derive the paradoxical conclusion,
that on occasion it has been 'Correct' for scientists to adhere to an
'incorrect' theory, even when offered what we now know to be
'correct'. For the philosophy of science of the nineteenth century,
 Scientific Inquiry: Problem-Solving on Artificial 01Jjects 127

such a conclusion would have been more than paradoxical; it would

have been anathema. So strong then was the commibnent to a faith
in a cumulative, infallible progress of scientific knowledge, that
erroneous opinions (either old ones, or established beliefs discovered
to be so) needed to be explained away by specially constructed
historical myths. Such was ~e fate of 'caloric', of 'phlogiston',
and of 'vitalism'; in each case it was shown that only a refusal of
certain scientists to open their eyes to the obvious kept the bad old
theory alive. What such a committnent did to the study of the
history ofscience needs no lengthy description here; historians must
still cope with the remnants of a crop of fantastic 'anticipations' of
scientific truths produced in that earlier period. IS
We can take this conclusion even further, and state the thesis that
the forward progress of science must necessarily be accomplished
largely by way of detours. For, again, it will only be the man of
genius who can sense, beyond the limitations of the evidence at his
command, what avenues of advance will be the most fruitful ones
II The most influential of these were the supposed 'precurson' of GaIiIeo'. dynamics
and kinematics in fourteenth-c:entury Paris, discovered by P. Duhan. See E. Moody,
'GaIi1eo, and his Precurson' in C.lileo Retlppr';setl, ed. C. Golino (University ofCalifomia
Press, 1966), pp. 23-43. Perhaps the most famous case of a theory whose mrrectne81 was
established only gradually, indirectly and inmmp1etely, is the heliocentric system of
Copernicus. For a century from its fint announcement, it suffered from its violation of
the principles ofthe only mhel'ent and accepted system ofphysics, that ofAristotle. The
two sorts of observational evidence that might support it, stellar para1Iu and effects on
falling bodies of the earth's rotation, muld not be produced until about three hundred
years after its announcement. For working astronomers, almost all its amstructioDS
muld be translated into models that did not involve the; motions of the earth. The one
exception, in the later sixteenth century, was the theory explaining the slow changes in
the elements ofthe orbits ofthe celestial bodies. Although this did win adherents to the
Copernican system, for it promised a solution to the pressing problems of calendrical
refonn, the problems were at the limits of the mmpetence of astronoJDaS, and plagued
by spurious data. GaIi1eo's own polemics were mainly against Aristote6an natun!
philosophy; his one sustained argument for the motious of the earth was his fallacious
and generally inmmprehensible theory of the tides. Thus we can say that, up to the early
seventeenth century, a judicious astronomer who had no metaphysical bias in IUs assess-
ment would return the opinion 'not proven 'on the Copernican system, and treat it as an
hypothesis.
Through the seventeenth century, both natural philosophy and astronomy provided
an increasingly powerful set ofindications that the Copernican theory was true in essence
if not in detail. Some of Kepler's technical achievements (as in the case of the problem
of planetary latitudes) depended on a hcliostatic system; the orbits of comets muld be
interpreted only on its basis; and the Aristotelian system fell out of favour. Finally,
Newton's great achievements in the prediction and explanation of the planetary motiODS,
which muld not be mmprehended except on the assumption of a rotating and orbiting
earth, provided 'moral certainty' to the heliostatic hypothesis. By this time, alllhe tech-
nicalities of Copernicus's astronomical system were utterly obsolete.
128 The Achievemmt ofScientific KnoT1Jledge
over the years after his own work. The others, working with the
imperfect materials available to them, proceed in the most natural
direction in their choice of problems and in their assessment of
controversial conclusions. That the path does get straightened out
eventually, or rather that a suCcessful path can be recognized in
retrospect, is an historical fact; for we have the body of genuine
scientific knowledge as wibless to the possibility ofsuch a corrective
process. How the process works is a matter which we will discuss
later at the appropriate point.
The artificial and fallible nature of the conclusions of solved
scientific problems has an importance beyond the improved his-
torical understanding of natural science. For the same holds true
in any field of thought and action where there is an explicit body
of theory in whose terms arguments arc cast and conclusions are
drawn. Since so much of the direction of the technical and social
aspects of our existence is now done by specialists trained in a
formal 'science' oftheir craft, an appreciation ofthe limitations ofthe
conclusions to scientific problems is as important for politics as
much as for epistemology.
Working on (I ProlJlem
Having sketched the 'anatomy' of a solved scientific problem, l
may now indicate some aspects of the 'physiology' of scientific
inquiry. The distinction of the different phases of work, relating to
the production of data, information, evidence, argument and con-
clusion, does not imply that these are successive, discrete and in-
dependent stages of every investigation of a problem. Although the
work on a problem tends to be dominated by different tasks at
different stages of its progress, there is an alternation among them,
and they also condition each other's plan and performance.
For example, we have already seen that the choice of the data
which is to be produced is influenced both by the sort of infor-
mation which is desired, and also by the special tools with which
the data and the information will be worked. In its tum, the informa-
tion is intended to function as evidence; and the evidence itself
depends on the particular form of the argument which will be
developed, and hence on the conclusion which one hopes to
derive. Thus in this work, as in Aristode's analysis of handi-
craft, the 'final' cause is the first, in the sense that it conditions
all the others. In the investigation of scientific problems, however,
 Scientiji& Inquiry: Problem-Solving on Artificial Objects 129

the conclusion cannot be completely known in advance. In such a

problem (as distinct from other $Orts), if one starts with a precise
question admitting a small range of well-defined answers, then it is
almost certain that the result will, as an enrichment of knowledge,
be nearly ttivial. I9 For in such cases, nothing new is gained by the
interaction with the external world; and the solved problem yields
little more than what was put into it, in the form ofthe first question
that could be fonnulated. In scientific work of high quality, one
must start with a general idea of the sort of conclusion that might be
achieved, and of the sort of evidence that might be produced to
support it, and then gradually develop a programme of detailed
work. 20 This is always tentative and 'experimental' at first, for
unexpected results at any point will modify the possibilities of the
use of the materials at later stages, and also require the provision of
new materials at earlier stages of development. Pitfalls can be
encountered at any point, and when they are discovered there must
be surveys of the damage, and a reorganization of the work for the
necessary repairs to the incomplete structure of the argument.
When the work on a problem is well under way, there is a hier-
archically ordered set of subsidiary problems under investigation,
each one concerned with the production or perfection ofacomponent
ofthe intended argument; perhaps new data or information, or some
particular chain of reasoning. These subsidiary problems are of a
19 In this point I am 1101 arguing that the answering of a precisely framed question is
necessarily a trivial undertaking. The testing of a hypothesis by a properly designed
experiment may be necessary for aucia1 evidence for a deep problem. The issue here is
how much work has gone into the development of the question, between the fint con-
ception of the problem and its refinement into its penultimate form. If the scientist has
learned nothing about the problem, or if there has been nothing to learn about it, once
he had the first partly clear idea of it, then it is hardly likely that there will be much in the
result from which anyone else can learn. The distinguished American physicist Karl
Compton put it succinctly: 'When I was directing the research work of students in my
days at Princeton University, ••• I always used to tell them that if the results of a thesis
problem could be foreseen at its beginning, it was not worth working at.' Q!toted in D.
s. Greenberg, Politics ofAmeri,,,,, S&im&e, p. I5S.
20 For historical scholarship, an illuminating account is given by E. H. Carr, Wiull is
History 1 (MacMillan, London, Ig6I), ch. I, 'The Historian and his Facts' (pp. 1-28).
Describing the interaction of the two activities of reading and reflection (either in the
mind or in writing drafts), which are mutually enriching and correcting, he says: 'Hyou
try to separate them, or give one priority over the other, you fall into one oftwo heresies.
Either you write scissors-and-paste history without meaning or significance, or you
write propaganda or historical fiction, and merely use the facts of the put to embroider a
kind ofwriting which has nothing to do with history' (p. 23). In this fashion Carr disposes
of some classical pseudo-problems of the philosophy of history, which arise from a c0n-
ception of historical research that is a caricature of its actual craft practice.
130 The A,hie'Oemmt of Scientific K,,01lJledge
different character from the principal problem under investigation.
They belong to a class which I shall later discuss as 'technical
problems', where the function of the required information or device
is known in advance; in this case, from the requirements ofthe main
problem. Such technical problems do not have the same freedom to
grow and evolve; and where they are subsidiary to a larger investi-
gation, they can be left to less experienced or less imaginative
workers. Of course, when a subsidiary problem turns out to be
incapable of solution to the standards of adequacy required for the
function its result will perform in the completed argument, that
part of the argument must be re-examined to see what other sorts
of results will be adequate there, and new subsidiary problems set
and investigated. It is in this complex fashion that a problem grows
through interaction of the scientist with his materials. The inspired
guesses and great illuminations which bring a problem into being
are only the beginning of the work; and the impersonal, demon-
sttative narrative of the published research report are only a stylized
record ofits completion. What lies between is demanding and subtle
craftsman's work on the very special objects of scientific knowledge.
But it is not an entirely undifferentiated set of intuitive procedures.
Using the categories of data, information, evidence, and argument,
one could construct a chart describing the state of the work on the
main and subsidiary problems of a scientific inquiry at any point of
its development. The organization of complex administrative and
engineering projects had been considerably helped by the use of
charts for flow-diagrams and 'critical path analysis'. A chart along
the lines suggested here would, at the very least, show by its com-
plexity the degree to which a scientific project was other than
straightforward routine; and a sequence of revisions of the chart
would provide a record of the evolution of the problem itself.
The similarity of structure between these other sorts of projects
and a scientific problem should not obscure one very deep difference.
Although the objects ofsuch practical or technical problems are also
artificial to some extent, they do not change their nature in the
course of the work. One of the things that makes scientific problem-
solving so uniquely subtle is that the very objects of the work evolve
as the work goes on, and in a fashion which is not predictable in
advance. For the discovery of new and unexpected properties of the
objects ofthe investigation entails a change in the objects themselves;
the objects described in the conclusion of a problem with genuine
 Scientific Inquiry: Pro1Jlnn-Solving on Artificial OlJjects 131

novelty are not those which existed when work on the problem began.
This is not a purely philosophical point; for in such problems, there
is a need for a frequent review, and a re-setting of subsidiary
problems, lest the work done at the earlier stages' of the project be
irrelevant to the needs of its conclusion.
Thus in a significant scientific problem, the objects of inquiry are
themselves plastic. They can remain unchanged throughout the
investigation only when the scientist is unusually prescient when
he sets the problem, or when the problem itself is banal. It is for
this reason that the investigation of real scientific problems cannot
be reduced to 'asking a question of Nature' with the expectation ofa
simple answer, or 'testing an hypothesis' to see whether it is ttue
or false. Even the description of 'normal science' as 'puzzle-solving'
canies the connotation that the puzzle is there in advance, and also
that it has a unique solution.:&1 Much of the routine work done by
scientific workers is such puzzle-solving, on essentially technical
problems. But good scientific work, even when it is not revolutionary,
is more demanding, and more interesting, than that.
The work on a particular problem is completed when an argu-
ment, meeting the accepted standards of adequacy, can be framed
and conclusions drawn. But this does not occur suddenly, as with
the dotting of the last i. Rather, the cyclic interaction of the various
materials of the problem decreases in intensity, the argument is
stabilized, the evidence becomes sufficiently strong and well-fitting,
and fewer lots of new information and data are required. The un-
expected results decrease in importance, pitfalls become negligible,
and there is a sense that the whole process is 'converging' towards
solution. But this does not always occur; and a warning sign of the
imminent failure of a problem is when the difficulties begin to
'diverge'; when the subsidiary problems called into being by new
difficulties become larger and more fundamental, and ever more
extensive modifications of the argument become necessary, with
increasing lots of fresh data being required for throwing into the
breaches. An important part ofthe craft skill ofa scientist is to sense
whether a problem is beginning to converge as early as it should, and
to detect signs of incipient divergence; and then to decide when to
abandon a doomed venture. It is the lack of such a skill that takes
21 These terms derive from T. Kuhn, Tile StnI&l"e of Scimtijie Revollllions, ch. 4t
'Normal Science as Puzzle-Solving'; the dichotomy between such 'normal' science and
'scientific revolutions' expressed in that work is, as Kuhn soon saw, somewhat too sharp.
132 The AclJieTJemmt ofScientific K1I01IJledge
beginning research students through to hopeless muddles, and on to
the final despairing struggle to salvage something of value from the
wrecbge of a failed problem.
'Stimtific Pro1J/nn' Defined
Having identified the constituents of a completed scientific
problem, by analogy with the four 'causes' of Aristotle, and also
having sketched the cycle of operations by which a problem is in-
vestigated and solved, we can now attempt a formal definition of the
term 'problem' itself. This term is becoming common, both in the
description of scientific work and in its philosophical analysis. The
concept requires some closer analysis, both for its philosophical use,
and for the distinction between 'scientific problems', which I am
discussing here, and some related, but different, sorts of problems
which I shall discuss later.
When can we say that a problem 'exists' ? We would like to be able
to distinguish between a 'problem' which is ready for investigation,
and less well-defined things, such as a sense that a certain sort of
difficulty, theoretical or practical, needs to be resolved; or that a
certain sort of work could or should be done. Any definition of the
instant of birth of a 'problem' out of a 'problem-situation' is bound
to be arbitrary, since all the components of a problem undergo con..
tinuous and possibly radical change during its investigation. Taking
our clue from Aristotle, we can say that the 'final' cause is the first
one; only when there is some specificatien of the new conclusion to
be drawn can we say that a problem exists. With this as the basis, all
the other constituents fit into place. The 'material' cause, the existing
stock of relevant scientific results as it is to be modified by the
solution of the problem, is presupposed in the statement of the
problem. Although the 'formal' cause, the argument from which the
conclusion will be derived, cannot be known in detail, the accepted
patterns of argument for the field can usually be presupposed. Of
course, in a deep investigation where this itself is modified, the
statement of the problem is correspondingly more difficult. But
something must be known of the 'efficient cause'; for a conjecture
or hypothesis which is not accompanied by a plan for its establish-
ment may be quite interesting but it is not something on which work
can be done.
Herein lies another reason why it is insufficient to characterize a
scientific problem simply as 'a question put to Nature', or as 'an
 Scientific /tllJuiry: ProlJlem Solving on Artificial OlJjeGts 133
hypothesis to be tested'. Judging such questions on purely internal
features, as their surprise, improbability, or organizing and unifying
power, can lead to utterly unrealistic accounts of the evaluation of
scientific problems. 2 :& The question must contain, in addition to its
implied"answer, some plan (implicit or explicit) for the attainment
of the answer. For the solution of genuine scientific problems is not
merely having bright or even brilliant ideas; these are empty unless
they are developed and enriched by the hard, complex and sophisti-
cated craft work of scientific inquiry. Unless there is some idea of
how the work will be done, there is no way of knowing whether the
solution can even be achieved; and in general the form that the
tentative solution takes will depend on the projected means of its
accomplishment. This is of course a matter of degree; as a limiting
case there is the handful ofclassic unsolved problems in mathematics.
But these are accepted as genuine problems either through the
naturalness of their objects, the empirical evidence for their truth,
or through the authority of their proposers.
The use ofthe term 'problem' to define the programme ofinquiry
goes back to antiquity in mathematics; although even there it has the
connotation of being able to do something rather than to know some-
thing. It has occasionally been used to give point to a precept of
method in other fields, the most famous example being Lord Acton's
dictum, 'study problems, not periods'. But it has tended to carry
with it the implication that solving a problem is a form of inquiry
which is both safer and less deep than something else which is being
abjured. It would be most unfortunate if this present discussion
were taken as evidence for the proposition that 'all the scientist does
is to solve problems' as if this were equivalent to doing crossword
22 Lest the very idea of'hypothesis' be lost when the so-called 'hypothetico-deductive
method' loses favour among philosophers of science, I should remark here that the
pattern of argument in statistical inference necessarily involves the report of the testing
or a prior hypothesis; and many pitfalls of interpretation can be encountered unless the
work actually proceeds by first framing the hypothesis and then getting the relevant data.
But a worthwhile hypothesis for statistical testing can only be set when the problem has
gone through its earlier stages of development, and is ready to be put in a final, crysta1-
lized form. The deep and complex problems of the validity of such real 'hypothetico-
deductive' arguments are little studied by the philosophers who have concerned them-
selves with this 'method' in the abstract. It is Doteworthy that the BritisA Jowul ftW tile
PllilosopAy ofStimtt, a leading journal in its field, has had just a single paper on the logic
ofstatistical inference in the twenty years of its publication. The exceptional paper which
proves my rule is I. Hacking, 'On the Foundation of Statistics', BritisA Jowul ftW tile
PhilosopAy of Stimtt, IS (1964), 1-27. See 'Index 19Scrlg6g Volumes 1-20', BritiiJI
Jowglfor 1M P/JilosopAy of Samet, 21 (1970), 21.
134 The AcAievemmt ofScientific Kno1lJledge
puzzles. For I am arguing that whatever a scientist does, it is best
conceived as the investigation (including both the creation and the
solution) of problems. We shall see that problems can vary in depth
from the trivial to the profound, and that when genuine scientific
knowledge comes to be, it is achieved through a complex social
endeavour, where the materiaIs embodied in the solution of one
problem are tested and transformed through their use in the in-
vestigation of subsequent problems.
We may now put our definition formally, and say that a scientific
problem is a statement (always partial and subject to evolution) of
new properties of the objects of inquiry, to be established as a con-
clusion to an adequate argument, in accordance with a plan (specified
to an appropriate degree) for its achievement. The naturalness of
this definition, for the social activity of science in the later twentieth
century, can be seen from its implicit use when applications are made
for research grants and contracts. The two components, statement
and plan, which we might call the 'final' and the 'efficient' causes of
the problem, are each necessary, and are joindy sufficient, for
specifying the project. On their basis, the judgements of value,
feasibility and cost can be made; and it is in terms of such judge-
ments that the decision is made for investment in the work. The
evolution ofthe problem in the course ofinquiry is a common occur-
rence; those who control investment in scientific work are aware
of this tendency to change, and they can also tell when one prob-
lem has turned into another.
With this definition of 'scientific problem', my analysis of the
cycle of the first, individual, phase of the achievement of scientific
knowledge is nearly complete; in a subsequent chapter I will discuss
the second, social, phase, in which the solved problem is accepted
as a research report and is then subjected to further wotk before
becoming a 'fact' and ultimately 'knowledge'. In this chapter we
have seen how the problem is solved by a conclusion being drawn
about the classes of intellectually constructed things and events
which are the objects of the investigation. The conclusion is the
outcome ofan argument, and is related to the external world through
the evidence embedded in the argument. This evidence derives
ultimately from the data, which is the sole point of contact with the
external world, through the information which is produced from it
through the application of tools. Throughout, the work involves the
use of craft skills, and the making of a variety of judgements; the
 Scientific Inquiry: Problem-Solving on Artificial Objects 135
character of both of these is conditioned by the artificiality of the
objects of the inquiry.

The Origins ofa Problem

This is an appropriate point to complete the description of the
cycle of investigation of a problem, and to describe how a problem
comes to be. The history of any inquiry begins with something that
is less than a problem; we may call it 'problem-situation'. By its
very nature, this is difficult to define. It may be considered as an
awareness that there is a question to be asked, without anyone being
able to frame the question successfully. Alternatively, it can be
regarded as a recognition that certain functions might be performed
by the solution of some problem as yet undefined. The variety of
such functions corresponds to the complexity of the frontier of
established results in a field. There may be phenomena calling out
for explanation; there may be a conflict between two schemes of
explanation; a given result might seem capable of being made more
deep or general; an established conclusion might appear to require
refutation or correction; the list can go on and on, and it would be
poindess to try to give an exhaustive formal account to describe
what is the particular craft practice of each field of inquiry.z3
23 Examples ofeach of these functions of a problem yet to be defined, are as follows.
The high-energy particle accelerators of the 19505 and 1960s produced a multiplicity of
phenomena; for the theoretical explanation of some of them, problems could be set and
solved; for others only problem-situations remained. The production of work by a heat-
engine could be explained, in the 18405, either by a 'fall' ofcaloric in temperature analo-
gous to that of water in height, or by the 'conversion' of sensible heat into work; the
conflict between these two set the problem-situation for William Thompson (later Lord
Kelvin). (I am indebted to D. S. L. Cardwell for this example). The problems of the
electtodynamics of moving bodies, and their apparent incoherence in failing to be
Galilean-invariant, set the problem-situation for the young Einstein. See G. Holton,
'Einstein, Michelson, and the "Crucial" Experiment', Isis, 60 (1969), 133-eJ7. Fourier
was put on the track of his successful theory ofthe conduction of heat in solids by a paper
of Biot on the steadY-6tate tempentures of a heated bar; Biot's physical model was
superficial and inadequate of its analysis, and Fourier immediately saw that it could and
must be deepened somehow. See I. Gratton-Guinness, 'Joseph Fourier and the
Revolution in Mathematical Physics', JounuU of tile ItUtitJItt of M.timu#i&s tUUl its
Appli,.tio"s, 5 (1969), 230-53; a book on Fourier's work by Gratton-Guinness and
myself is forthcoming. Finally, it is likely that Davy's antipathy to the 'French'
system of chemical nomenclature posed a problem-situation of refutation; this crystal-
lized into the problem of demonstrating that the 'oxy-muriatic base' of muriatic acid
was, by Lavoisier's own principles, an element. The thick green gas was subjected to
every possible process of decomposition, and resisted; hence Davy named it 'chlorine',
and destroyed the foundation of the nomenclature based on 'oxygme' being the prin-
ciple both of burning and of acidification (from which its name was constructed).
 136 The Achievement ofScientific Kno1lJledge
The problem-situation will persist as long as there is a recognition
of the function to be performed, without a successful framing of the
problem which will enable its performance. This phase ofthe 'poten-
tial' existence of a problem may be very brief, but for really deep
problems, the bringing of the problem itself into existence can be a
long, arduous, and hazardous operation.
Relating the problem-situation to the Aristotelian scheme of the
task ofinvestigating a problem, we notice that the task itself, defined
by its goal of the establishment of certain new properties of the
objects of inquiry, depends for its existence on a prior 'end': the
function defines the problem-situation. In this there is a point of
similarity between scientific problems and technical ones; but in
the latter case the function remains dominant throughout the inquiry,
while here it can easily be displaced if the investigation of the
problem reveals new possibilities in the material.
Although the problem-situation is in a less specified state than the
problem to which it may give rise, it is already a very artificial thing.
The very existence of a problem-situation presupposes a mattix of
technical materials: existing information (with the intellectual
objects it describes), tools, and a body of methods including criteria
of adequacy and value. For in the absence of such a mattix of
technical materials, a genuine problem could never come into
existence, and the decisions on investigating it, and on shaping it
during the work, would have no foundation.
Because of the unspecified character of a problem-situation, its
development does not require the precise thinking and concentra-
tion ofattention that a fully developed problem demands. Hence the
perception of what is needed to crystallize it into a well-defined
problem need not come as the result of straightforward toil. There
are well-attested cases ofmathematicians suddenly having a (correct)
insight into the solution of a difficult problem, usually during a
period of rest or distraction from intensive and unsuccessful work. 24
In natural science, the classic cases of crystallization ofa problem on
perception of a commonplace event tend to be apocryphal, for the
problem-situation itself is already so technical that the 'falling
apple' will provide insights relating to it only indirectly. And in the
course of an investigation leading to a great discovery, the problem-
situation itself will tend to be fluid, and the identification in retro-
24 See J. Hadamard, Tile Psy,IJoloD ofInvmtioll in the M.t!Jnuti,.1 Field (Princeton
University Press and London, Oxford University Press, 1949).
 Scientific Inquiry: Problem-Solving on Artificial Objects 137
spect of that most crucial insight will tend to be conditioned both by
later events and by fashion in theory of discovery. But without a
problem-situation, a crucial phenomenon, whether commonplace or
artificial~ would go unremarked either as ordinary or as a stray
anomaly. This is the basis for the maxim, 'chance favours the pre-
pared mind'; and the more trivial the crucial phenomenon, the more
certain it is that the observer had been preparing for it intensively.
The mind must be well-prepared for yet another reason; for the
phenomena, as they occur, will have none of the reliability and
definition of the data which are later produced in the course of the
investigation of the problem. They are as likely to lead to pitfalls as
they are to show the correct path. The phenomenon of the limited
working height of lift-pumps misled GaliIeo; for he related it to the
problem of cohesion in solid bodies, on which he was building a
sort of solid-state physics, and in doing so he was led away from the
correct explanation of the hydrostatic phenomena previously ex-
plained by the principle of 'Nature's abhorrence of a vacuum'. Z5
Following on the perception of a crucial phenomenon comes the
initial insight into the problem and its solution. At this point one
may speak of the problem being born. This is the most exciting and
also the most dangerous time for the scientist. Q!tite suddenly all,
or nearly all, the pieces of the puzzle fall into place. The conclusion
to be drawn takes shape, becomes real, and awaits only the per-
formance of the routine detailed work to be acceptable as genuine.
At this point in the work every scientist of real originality and com-
mitment is filled with a 'sober drunkenness' ;Z6 the experience is as
close to the mystical as any other in our culture. The dangers arise
because this initial insight will determine the form of the full
specification of the problem, and will include error along with truth.
2S In the T1lJO NeJIJ Snmees, the lift-pump phenomenon is introduced after the dis-
cussion of cohesion in solids with which the book begins (Edizione Nazionale, p. 64).
The importance of the awareness of the problenHituation for a recognition of
phenomena is shown by C. SUsskind, 'Observations of Electromagnetic Wave Radiation
before Hertz', Isis, 55 (1964), 32-42. There were many observations of the reception of
electric signals at a distance, but these were unsystematic and easily dismissed by leading
scientists.
26 This term is borrowed from L. Edelstein 'Recent Trends in the Interpretation of
Ancient Science',:/ounuJ1 oftu History ofIdea, 13 (1952, 573-604; reprinted in Roots of
Snmtiji& TJuJu,IJI, ed. P. Wiener (Basic Books, New York, 1957), pp. 90-121. He quotes
Ptolemy: 'I know that I am mortal, the creature of a day; but when I search into the
multitudinous revolving spirals of the stars, my feet no longer rest on the earth, but,
standing by Zeus himself, 1 take my fill of ambrosia, the food of the gods', p. 583
(p. 100 of the Wiener volume).
 138 The A,hievement ofS'imtijie Kn01lJledge
Days or years can later be wasted in an attempt to confirm what
once seemed blindingly obvious. It is likely that the braziers which
GaliIeo saw swinging regularly In a church had this hidden danger;
for he never ceased to believe, and doubdess spent much time
trying to prove, that the period ofa simple pendulum is independent
of its amplitude. Z7 Worse yet, it takes superhuman coolness for a
scientist later to recognize that his great insight was misleading, and
in effect to amputate part of his life in rejecting it. The deeper and
more powerful the initial insight, the greater the force of personal
identification with the problem it defines. This can last for the rest
of a man's life, long after the initial insight has been whittled down
by destructive criticism, or bypassed by the progress of the field.
From such situations are personal tragedies made. They also pose
the most fascinating problems for historians of science, for the
initial insight is frequendy buried deep in the body of the mature
work which finally results from it; yet to interpret that work his-
torically it is necessary to reconstruct it from its beginnings.
The setting ofthe problem is not completed by the achievement of
the initial insight. This must be refined in the light of what is avail-
able from the existing results in the field, including information,
explanations and tools. Also, there must be a judgement of the
degree of specification which is appropriate to the problem at this
stage. Too complete a specification can result in the waste ofeffort on
incorrect lines ofattack; too little specification can lead to an aimless
wandering around the field. Above all, the scientist must examine
his nascent problem, and make the judgements of value, feasibility,
and cost; and only then decide whether to gamble his time and
(usually) someone else's money, on its investigation.
Thus in the early stages of the creation of a problem, routine
craft operations are less significant than imagination and judgement;
and so this phase of the work will be strongly influenced by the
personal style of the scientist. Moreover, at this crucial early stage,
the character of the work required, and the appropriate attitudes,
change with great rapidity. At first one must be able to think in an
analogical and associative fashion, to perceive the crucial phenomena
for what they are; then to proceed boldly and speculatively to the
initial insight; but soon after to be rigorously self-critical in the
definition and assessment of the problem. It might be tempting to
imagine these different functions split up among members ofa team,
27 See T1lJO NeJIJ Snmeel, Edizione Nazionale, pp. 139-41.
 Scientific Inquiry: Problem-Solving 011 Artificial Objects 139
benefiting from each other's strengths and protecting ag;Unst their
weaknesses. But since deeply novel ideas cannot be communicated
until they have received a fair degree of specification, the basic work
of creation, with its special hazards and pitfalls, must rest with an
individual. This is why the achievement of great discoveries in
science is not a matter of perceiving the obvious, or still less of luck,
but is reserved to those with greatness of intellect and of at least
some aspect of character.
This dramatic picture of the investigation of problems may tend
to convey the misleading impression that all scientific work is an
heroic endeavour, as in a style of propaganda for science which was
once dominant but which is now becoming rapidly obsolete. I shall
later discuss the problems of mass-produetion in science; but for
now it will suffice to indicate the diversity among problems, in
respect of their degree of elaboration and their challenge and diffi-
culty. It is easy to see that the evolution ofa problem, and the change
of the objects of inquiry during its investigation, will tend to be
most rapid and dramatic in those fields where the argument is most
highly elaborated, and where the production of data from experience
is a relatively small (although perhaps crucial) part of the work.
These are the 'theoretical' sciences; at the other extreme are the
'descriptive' ones, where the data is largely found rather than manu-
factured, and where a mass of data provides evidence in a simple
argument. In the latter sort of field one might seem to be 'fact-
collecting' were it not for the presence ofpitfalls at every stage ofthe
work in all but the most routine of problems. Even though great
and dramatic advances can be made in descriptive sciences, the pace
is slower, and the style of original work is different: breadth of
experience and maturity of oudook weigh more heavily than in the
abstract, argumentative fields. In such fields, one may consider a
scientist's work as consisting of a large multitude of very small
subsidiarJ problems, all of them guided and controlled by the
gradual evolution of a leading, significant problem.
But the real division among fields in respect of the excitement and
hazards of worthwhile work does not lie along this axis. Rather, it
relates to the proportion of the necessary work in the field which
consists of non-trivial 'technical' problems. In many fields which are
by no means descriptive, there is a need for large quantities of
reliable information, either produced to order for a particular
problem, or made publicly available in handbooks and digests. The
140 The A,hitvemmt ofS'ientijie Kno1lJledge
production ofthis information requires sophistication, diligence, and
care; for its objects are artificial and each new project encounters new
pitfalls. So the work demands something more than a technician's
ttaining; in the social classification, it is scientific. Yet when pursued
by the unadventurous, it can become a routine, never completely
repetitive and boring, but a straightforward craftsman's job all the
same. The style of work which is appropriate for such research
would be of little use in highly theoretical or speculative fields; but
the great rank and file of scientific workers spend their careers in
setting and solving such high-level technical problems. Whether
one wishes to describe such work as 'science' is a matter of choice;
but its presence, as an essential ingredient of the industrialized
science of the present, cannot be ignored.

Craftsman, Te'hnician, Scientist

My final application of the analysis of scientific work as problem-
solving on a world of artificial objects will be to develop the distinc-
tion between the work of a scientist, a technician, and a craftsman.
Through this we will come to an improved historical perspective
on the picture of the natural world in which our contemporary
natural science operates. I have already argued that every scientist
must be a craftsman in some sense, and he must be a technician
too. But the objects of the work of a ttaditional craftsman and of a
technician, and the tasks they accomplish, are each distinct; and we
can define the differences in terms of the special character of the in-
tellectually constructed objects of scientific knowledge.
Craftsman's work is done with particular objects, which may be
material or intellectual constructs, or a mixture of the two: and the
operator must know them in all their particularity. Their properties
and behaviour cannot be fully specified in a formal list; in fact, no
explicit description can do more than give the first simple elements
of their properties. Hence the operator's knowledge of them must be
'intuitive', or of the sort described by Polanyi as 'tacit'. z8 It cannot
be learned from books, but from experience, derived from a teacher
by precept and imitation, and supplemented by the personal
experience of the operator himself. Such a craftsman's knowledge of
his objects is necessary for any sort of scientific work; even in pure
mathematics, where the objects of the work are purely intellectual
21 See M. Polanyi, Perstmll1 Ku",ktJ,e, pp. 87-95.
 Scientific /nlJu;'y~· Problem-Solving on Artificial Objects 141

creations, the properties of the objects which are known from

established results are not sufficiendy particular and subde to guide
the work of constructing an argument for a new proof.
The craft character of the scientist's knowledge of his objects is
even more marked in the case of the basic, elementary concepts
which are learned at an early stage ofttaining. For the teacher cannot
exhibit the classes of things and events as parts of a formal and
coherent structure; rather, he must try to make them plausible, and
impart some knowledge of their properties through the student's
successful manipulations with them. They are thus learned in-
formally, by imitation and precept; and they become so deeply
embedded in the mature scientist's picture of the world that they
eventually seem entirely 'natural'. One cannot imagine an intelligent
adult who cannot count; and so the scientist (or the teacher of
science) finds it difficult to accept that some intelligent and educated
adults cannot manipulate with such basic concepts as 'mass' and
'acceleration'. Embarrassment can arise when this craft knowledge
of the artificial objects of science is assumed by its possessors to be
the total possible knowledge; probing questions by awkward students
then threaten to expose a vacuity of incomprehension beyond the
fa~de ofconfident manipulative knowledge. Of course, at some stage
ofttaining, the concealment ofthe artificiality ofthe objects becomes
impossible, and also unnecessary. Thenceforward, the student con-
sciously operates explicidy in a world of intellectually constructed
objects, but he is rarely reminded (and hence nearly as rarely learns)
that the 'natural', 'elementary' objects he learned in his youth are of
the same sort.
Even when the objects of a field are recognized as constructs, the
craftsman's familiarity with them which scientists achieve in the
course of their training has a conservative effect. For the imaginative
work of perceiving new phenomena and setting new problems will
tend to be done in terms of what is already intuitively known. It is
for this reason that the research schools founded by great men tend
to lose their originality; and it also explains why deep innovations
may arise from the new perceptions of scientists outside a leading
school, or outside the field altogether, or from young, self-ttained
men. In such cases, the innovator will need enormous courage and
talent to overcome the pitfalls in the work, which would have been
at least partly recognizable to the 'properly' ttained men, had they
escaped from their familiar world into the partly new one. Hence the
 142 The AcAiewmmt ofScientific K1UJlIJledge
chances for success in such cases are small; only a few fortunate and
talented outsiders can crash into a field and make it partly their own.
If they do succeed, then of course they are retrospectively admitted
to respectability, and may in their turn contribute to a new ortho-
doxy of common sense objects.
Given all these similarities, traditional craft work differs from that
of scientific inquiry both in its objects and in its tasks. Its tasks are
the making of particular things and the control of particular events;
and these particular things and events do not derive their existence
from an elaborated, formal framework ofknowledge of intellectually
constructed classes of things and events. Most of traditional craft
work does not require formal literacy, although good craft work
does require intelligence: each new task presents new difficulties
and new possibilities.
The work ofa technician is also with particular things and events:
but some of the properties which he must understand and control
require the intellectually constructed classes of,scientific' knowledge
for their explanation or even for their description. Hence the training
of a technician will be radically different from that of a craftsman.
The apprentice craftsman can soon learn to perform simple manipu-
lations on the objects of his mature work, either learning by imita-
tion and precept from a master, or by making things himself which,
though lacking finish and sophistication, are genuine objects of the
craft. On the other hand, the technician in training must master a
considerable body of knowledge of an absttact, scientific character
before he can manipulate or even recognize his objects. This back-
ground scientific knowledge will, in its teaching, be designed quite
explicidy to serve as tools, and so it will lack the elaboration and
attempted rigour which is given to the same material when it is taught
to intending scientists. Yet since the technician's work will generally
be confined to the manipulation of a small set of objects, he must
also be given a thorough cr&ft apprenticeship for developing the
necessary skills. The proper blending of 'craft' and 'scientific'
elements in technical education is difficult even as a pedagogical
problem, but it is further complicated by considerations of status.
For rewards of 'brain' work are so significantly higher than those of
'hand' work that there is a constant tendency to elevate a particular
body of technique towards the status of an applied science. This is
frequendy justified by the introduction of new and more sophisti-
cated tools; but it can have the result of turning out people whose
 Scientific Inquiry: Problem-Solving on Artificial O~jects 143
training leaves them unprepared, psychologically and technically,
for the only jobs open to them.
Turning now to the scientist, we see the essential difference
between his work and those of the other two. For although he must
have a craftsman's knowledge of his objects and his tools, and do
technician's work as well, his real work is quite different in character.
For unless he can successfully set, investigate, and solve problems,
drawing conclusions about classes of things and events and not
merely manipulating particular samples, his tide is inappropriate.
Because scientific work generally requires a greater capacity for
absttact thinking, and a more rigorous training, a career in such
work tends to be restricted to those who have absorbed education
from early years. This naturally gives it an enhanced social status
over technician's work; but it is a mistake to think that it is essenti-
ally of a higher order of intellectual endeavour. Scientific work can
be done well or badly; and it is quite possible for the most intelligent
and competent person in a research laboratory to be the senior
technician, who mayor may not be aware of the deficiencies ofthose
whom he serves.
This threefold classification is far from absolute, and two points
where it breaks down are worth special notice. The most common
case in the modem world is where the distinction between scientist
and technician becomes blurred. Many scientists are employed on
problems of such a routine character, where the work is so sttaight-
forward and the classes of things and events are modified so little by
the new knowledge gained, that the 'drawing of conclusions' is
little more than the description of certain properties of quite reliable
representatives of the general classes under investigation. In such
cases the work is hardly different from that of a technician as I have
defined it; and such work is done by men trained as 'scientists'
rather than as 'technicians' only because of the customary difference
between the two sorts of training in depth and rigour. The distinc-
tion is also blurred from the other end, in the many cases where the
accomplishment of a particular technical task involves such complex
and sophisticated procedures that the work is in effect the drawing
of new conclusions about intellectually constructed classes of things
and events. Both types of borderline case occur on a large scale in
the industrial applications of science; the terms 'research worker'
and 'technologist' give a fair indication of the quality of endeavour
required in the two cases.
144 The Achievement ofScimtijie Kno1lJledge
The Crll[tStlllln's Cosmos
A less familiar exceptional case arises in ttaditional crafts, which
seem to involve no more than manual skills learned by imitation.
One will find, on close inquiry, that no such activity can be entirely
free oftheory; the craft knowledge must always include an 'explana-
tory' component, if nothing else than to guide the work when new
situations arise•. It does not matter that such 'theories' may be (and,
over the course of the centuries, usually were) incorrect as scientific
explanations of the properties of the materials; without them, there
could be no successful craft work at all. But the matter does not
always end there. Over the millennia of human experience, the basic
productive crafts, which are now being extinguished by modem
science and industry, developed very highly articulated theoretical
structures. The operators considered themselves as engaged in
something far more complex and rich than mere manual activities.
We could say that they were being technicians in our sense, perhaps
even scientists, except that they did not so much draw conclusions
as enter into a dialogue with their objects.
This mental and spiritual aspect of craft work has long been a
puzzle to anthropologists and archaeologists, for the 'magical' com-
ponent of an otherwise sound technique seemed so patendy irrele-
vant and absurd. Z9 Indeed, since the same material objects can be
created by purely manual operations, one might simply dismiss all
the magical theory as aberrant, retrograde superstition. But this
methodological criticism derives from.. particular view of reality.
If one is quite sure that no personal being resides or participates in a
particular artifact, then the copy made by the self-taught anthro-
pologist is the same as the original; but otherwise, the difference
between the two will be that between the dead and the living. And
it is undeniable that the 'magical' craftsman had a spiritual ex-
perience as part ofhis work, even ifit was illusory; while the modem
scientific copyist does not.
or course, it is obvious common sense to us that the craftsman's
material objects are devoid ofspirit. But this common sense is a very
recent and specialized cultural product of European civiIization;
a, v. Gordon CWde wrestled with this problem, which has destructive consequences
(or the ideology of 'rationality' and 'progress', in the tenth Frazer Lecture, M.gi"
c"./tsrutUJlip MIll Srimee (Liverpool University Press, 1950). The same phenomenon
bas been made the basis for an analysis of the modem ideas of 'progress' as the successor
to the magicaklchemical vision, by M. Eliadej see Tile Forge."'" tile Cnm1Jle (Rider,
London, 1962).
 Scientific Inquiry: Problem-Solving on Artificial Objects 145
and its triumphs over the past few centuries in the creation of
modern science and technology are now revealing their dangers for
human society and for the very continuation of life on this planet.
We will understand ourselves, and our contemporary problems,
better if we appreciate the world of the traditional craftsman as
utterly different from our own, perhaps quite wrong, but coherent
in itself and with its own value for the proper interaction of our
species with the rest of nature. The recent growth of movements
rejecting runaway technology and the culture with which it is
associated is a reminder that our 'scientific' world-view is not
necessarily the whole answer.
 5
METHODS

HAVING described the peculiar character of the activity of science,

as a special sort of craft work operating on intellectually constructed
objects, we can proceed to an analysis of the judgements by which
that work is governed. From this discussion we will achieve a better
understanding of the ways in which the social aspects of scientific
inquiry condition the work of individuals. We will then be better
able to analyse achieved scientific knowledge as the product of a
social endeavour extending through time, and also to attain insights
into the delicate questions of the health of scientific communities.
I have already mentioned some of the judgements which are
necessarily involved in the investigation of scientific problems,
beginning with the basic judgement of the soundness of a set of
data. These individual aets of judgement do not derive solely from
private intuitions of the scientist; rather they are based on a body of
principles and precepts, social in their origin and transmission,
without which no scientific work can be done. I shall use the term
'methods' for such principles and precepts, which (through their
interpretation and application in particular situations) guide and
control the work of scientific inquiry. Such a corpus of 'methods'
has a special character, which may seem paradoxical and at variance
with the objects and results of the activity it governs. For methods
cannot be established 'scientifically', through arguments resting on
controlled experience; this is partly because there is no simple test
of the 'correctness' of a particular method, and even more because
the principles and precepts are incapable of a fully explicit, public
statement. The body of methods associated with any field is partly
informal and even tacit; it is not transmitted by the same public
channel of communication as scientific results, but through an
informal, interpersonal channel. Hence the testing, criticism, and
improvement of the methods for a field must proceed by means
 MetAods 147
quite different from those applicable to its scientific results; and in
this aspect of scientific inquiry, the character of the community
engaged in the work is thus crucial for the nature and quality of its
achievements.
There is no doubt that without an appropriate 'method', in some
sense of the term, scientific work is impossible. A trained scientist
can instantly identify the traces of the bungling amateur, or the
crank, by the absence of 'method' revealed in a report of his work.
Also, there is a rough distinction in practice between 'methods' in
the plural, referring to more detailed techniques of work, 'method'
in the singular, referring to more general principles, and finally
'methodology' which is, or should be, reflection on the fonner
categories or their equivalents. Yet, while in some fields of science
the practitioners are deeply concerned (sometimes even obsessed)
with 'method' in one or all of these senses, in other fields of a
similar character the whole subject is complacendy ignored. There
is even a division and confusion among 'methodologists' themselves,
concerning the relations of their analyses and conclusions to the
work actually done in the disciplines they are discussing. One un-
fortunate consequence of the apparendy abstract and self-contained
character of methodological debates is that working scientists with
a philosophical concern will sometimes, after sampling such dis-
cussions, dismiss the whole subject as irrelevant nonsense. 1
The root of the difficulty in any discussion of method is that it
involves an attempt to render explicit that which is largely tacit.
For the achievement of significant new scientific knowledge is a
creative activity, involving intellectual work that is both bold and
subde. No machine that has been conceived, let alone constructed,
is capable ofsuch work. Z Nor can there be a 'science ofscience' which
I Thus L. K. Nash, Tile N.twe of IIIe Nlltw.l Snmees, pp. 320fF. The section
entided 'Scientific Method, Shadow and Substance' begins: 'We never credit the
possibility of a definable Literary Method, practice of which brings within reach of
mediocrity the production of works of genius. Yet at least four reasons may encourage
even master scientists, and others, to believe possible a communicable Scientific
Method with some such magic power.' With Polanyi, Nash sees science IS craft work;
and he also knows of the manifold ways in which scientists can delude themselves. See
also P. W. Bridgman, Reflections of. PIJysicist (philosophical Libnry, New York, 1950):
'I like to say that there is no scientific method IS such, but rather only the free and utmost
use of intelligence. In certain fields of application, such IS the so-called natural sciences,
the free and utmost use of intelligence particuIarises itself into what is popularly called
the scientific method' (p. 278).
:I This is not to say that machines cannot do any work which, when done by mea, is
considered 'original research'. I am informed by Professor F. A. E. PiraDi that a com-
 148 The AcAievemmt ofScientific Kno1lJletlge
can produce manuals teaching how to do original research in ten
easy lessons. Hence any explicit analysis of scientific inquiry must
be incomplete, at best a schematic anatomy representing a complex
physiology. This limitation applies to the present discussion as to
all others; but we can hope that by recognizing its limits, and sep-
arating off the unanalysable from that which is partly analysable,
we may escape some of the pitfalls which have beset earlier discussion
of the problems of method.
We have already discussed 'methods' of one sort, in connection
with the techniques of using tools. Such methods are the most
similar to those which govern handicraft work: they can be learned,
at the elementary level at least, in terms of particular precepts
governing particular operations. Also, in considering scientific in-
quiry as a sort of craft work, we saw that a master-aaftsman of
science has a body of methods of his own, which define his personal
style of work. But scientific inquiry differs from handicraft work in
using a body of methods which are sophisticated and subde, but
which are a social possession, not restricted to the private craft
wisdom of a master. We encountered examples of such methods
when we discussed the necessary judgements of the materials pro-
duced in each phase ofthe investigation ofa problem, as the 'sound-
ness' of data, and the 'relevance' and 'reliability' of information.
Such judgements are of adequacy; an equally important set, in-
voked at different phases of the work, are those ofvalue. An analysis
of these two sets of conttolling judgements will provide materials
for a later discussion of the philosophical problems of the nature of
achieved scientific knowledge, in relation to the social activity through
which it comes to be. Also, by appreciating such judgements, and
the criteria on which they are based, as methods, we can examine
the status of 'methodology', and appreciate some of its special
problems as a field of inquiry. Finally, we shall consider what sort
of knowledge is constituted by the body of methods which govern
scientific inquiry, and contrast it, and its channel of communica-
tion, to the results of scientific inquiry.
AtlefJflll&Y: the LaGi ofCertainty
In our earlier analysis of scientific inquiry, we got as far as the
completion of an individual research report. At that stage the
pater propam aiIII whereby ....ble M.Sc. dissertatioDs in eauiD fields or applied
matbematicl can be produced with only minimal human intervention.
 Methods 149
product is a solved problem; the establishment, in the conclusions
of an argument, of new properties of the classes of intellectually
constructed things and events which constitute the objects of the
inquiry. Up to that point it is inappropriate even to speak of'faets',
to say nothing of 'knowledge' or 'truth'; these categories belong to
the later, social phase of the activity of science. A scientific problem,
unlike a textbook exercise, carries with it no guarantee that there
exists a 'correct' solution against which those actually achieved can
be tested. When this lack of certainty is recognized as significant, it
entails some very paradoxical and sometimes troubling conclusions.
For the certainty and objectivity of knowledge about the world of
nature, as proclaimed in an old and powerful tradition in the
ideology of natural science, are called into question. The individual
scientist can achieve no more than an adequate solution to his
problems; and the criteria of adequacy are set by his scientific
community, not by Nature itself. To be sure, in a healthy and
matured field these criteria, as those of value, will be set in the light
of successful experience of penetrating into the natural world; but
there is no objective, certain or 'scientific' method for setting or
testing them. It still remains possible that certain and objective
knowledge can be attained by a community of scientists developing
and testing it over generations; and we shall discuss this in detail
later. But this then amounts to veritas tempo," filia, 3 and is the
opposite of what has generally been conceived as the process of
scientific discovery. Indeed, a reflection on criteria of adequacy,
their importance and their origins, can lead to the conclusion that
the differences between work in the most matured of the natural
sciences, and in the most fledgling of the human and social sciences
are of degree rather than of kind.
These paradoxes are very real, and will be discussed more closely
later. In brief, we may say that they result from the peculiar
character of scientific disciplines, which is most easily discerned in
the more 'theoretical' ones. That is, in them we have demonstrative
disciplines whose conclusions cannot be certain. This uncertainty
has generally been studied only in the context of abstract epistemo-
logical debate. The processes whereby relative and provisional
3. 'Truth is the daughter of time'_ classical Latin epigram. This may have been the
source for the tide of Francis Bacon's unpublished essay 'Temporis partus masculus',
'The Masculine Birth of Time' of 1605; see B. Farrington, Tile PIliIoSIJp1J7 of F,-m
BlUtm (Liverpool University Press, 1964).
ISO The Achievement ofScientific Knowledge
certainty emerges from the social activity of scientific inquiry had
not been a subject of close analysis until very recently. Also, since
the earlier defenders of natural science were usually concerned to
establish its credentials to the greatest degree possible, these radical
flaws in the structure of scientific knowledge have not become
widely known, either among those in other disciplines who would
emulate its successes, or among those who teach it to the younger
generation.
The terms 'deduce', 'verify', and 'prove' are still so widely used
in science, especially in teaching, that it is worthwhile to give a
brief resume of the inconclusiveness of the basic patterns of argu-
ment used in natural science. The simplest and most common
situation in the teaching of a developed experimental science is
where a general theory is elaborated, experimentally testable con-
sequences are deduced, and the relevant experiments are then
performed. If the particular results agree with the predictions, the
student is satisfied that he has 'confirmed', 'verified', or even
'proved' the theory. But in this stark and simple form, nothing
whatever has been established. It is not merely that a particular
collection of data can be interpreted as information in a variety of
ways, so that five points which lie reasonably close to a straight line
can (in the abstract) equally well be interpreted as lying on a poly-
nomial of the fourth degree. Even if the information is accepted as
being equivalent to a deduction from the theory, the 'support' it
lends to the argument is, in strict logical terms, nil. For this pattern
of argument is a fallacy known since ancient times as 'affirming the
consequent'. On this simple pattern, one can 'verify' the hypothesis
that the moon is made of mouldy cheese. One need only deduce
that it would then have spots, and then establish that the predicted
spots do exist. In descriptive sciences, the statements which are
'verified' are generally those of the properties of members of a class.
A sample is studied, and if the claimed properties can be identified,
the general assertion is passed as true. But all that can be claimed
logically is that one more confirming instance has been found; and
no matter how many more are accumulated, they cannot offer the
slightest ground, in terms of any valid pattern of argument, for
expecting future experience to be the same.
In both of these patterns of argument, negative or 'disconfirming'
particular experiences seem to have a stronger logical force. For if
the consequent of a particular hypothesis turns out to be false, then
 Methods lSI

the hypothesis itself is certainly false. Thus, although the deductive

pattern of argument is quite incapable of telling us when our
theories are correct, we seem to have at least the gloomy consolation
of knowing certainly when they are false.• Unfortunately, the in-
ference back from a disconfirming instance works only in the case
of very simple arguments; if more than one hypothesis is involved
in an argument (as is nearly always the case in science), we can
learn nothing more than that they cannot all be true together.
Similarly, an assertion of particular properties of a class of things
is not simply overthrown when a single contrary instance appears.
The original assertion can be defended by a slight redefinition of
the class, so that the offending sample is then excluded; or the
sample can be dismissed as one of those 'anomalous cases' which
abound in any detailed study of the workings of nature. 5
Even this slender link with certainty is lost when probability
statements are contained in an argument or a conclusion. For no
set of samples can formally confirm or disconfirm a predicted
probability distribution. To establish their relevance, one must in-
voke an extra probabilistic theory relating to the mode of production
of the properties of the samples, then estimate the probability that
the deviation between results and prediction arises from 'chance',
and finally decide what is a 'significant' level of odds against a
chance deviation in this case. The special tools, assumptions, and
judgements involved in this sort of 'testing' ensure that it will
remain very far indeed from the realm of certainty.
Philosophers of science have attempted, with some success, to
provide a rationale for the different basic patterns of argument,
showing why it is reasonable for an intelligent person to place
reliance on them. A common technique is to exhibit ancillary ~
sumptions, which when added to the bare argument do make it
t
4 Karl Popper is considered with some justification as the 'philosopher offalsification •
But as his ideas on this have developed over the decades, the simple docttine ascribed
to him has been refined nearly out of existence. For a discussion of the vuious 'Pop-
perian' positions, see I. Lakatos, 'Criticism and the Methodology of Scientific Research
Programmes', Pr«tttli",s oftile Anstott""" SO&iet7, fig (1g68), 149-86.
5 Michael Polanyi describes how the theory of electrolytic dissodation proposed by
Arrhenius was accepted for several decades, in spite of applying only to weak electr0-
lytes; the sttong electrolytes, as common salt or sulphuric acid, were amsidered IS
'anomalies'; see PersOMI K,,011Jlttl,t, pp. 293-3. In his classic study of the Euler
Polyhedron Theorem, I. Lakatos shows how 'monster-barring' strategies are also
applied in pure mathematics. See his 'Proofs and Refutations', BritisIJ JounuU for tile
PAilosopAy ofSlimet, 14 (1963-4), 1-25, 120-39, ZZI-43, zg6-3.p.
15 2 The Achievement of Scientific Knowledge
valid. But as these philosophical arguments become more refined
and sophisticated, they drift further and further from the practice
of science. They remain as important studies in epistemology, but
they offer little illumination to a scientist on the reasons why the
complex patterns of argument he actually uses seem to work and
are accepted as adequate by his community.
For what happens in any practical situation is that the basic
invalid patterns of argument are not so much supplemented by
extra assumptions as enriched by particular tests and conttols
appropriate to the materials of the problem and its characteristic
pitfalls. For example, a deductive argument may fail the test of
adequacy in a particular field, unless a set of known likely alternative
hypotheses for the same tested conclusion are discussed and
eliminated. This is no logical protection against the infinity of other
hypotheses which might yield the same conclusions, but it is an
insurance against at least the more obvious pitfalls. In an inductive
argument, the quality of the sample can be scrutinized, and special-
ized arguments used to show that it is likely to be a good repre-
sentative of its population. The making of inferences of causal
connections between members of two classes can be tested through
supplementary investigations, in the manner laid down in Mill's
'canons of induction'. In an elaborated argument of any sort, a
piece of evidence that has a significant load put on it may be re-
quired to be tested for its sttength by a particularly close examination
of the conditions of the production of the data and information
from which it is derived. None of these tests provide certainty, and
there is not even a certainty that a particular test is the most appro-
priate one for the problem being investigated. But they are the best
that can be obtained, for the work of seeking knowledge about the
external world; and a study of their special character will provide
an understanding of the ways in which that knowledge can, but need
not, approach certainty.

Judgements of Adequacy
Each component of the argument of a solved problem, either an
inference-link or a piece of evidence, can be no more than adequate
to its function in the total sttucture. And what is 'adequate' will
depend not merely on its context in the problem, but on the general
criteria of adequacy for the class of such problems imposed by the
community. In the beginning of this part I gave an example of the
 Methods I S3
necessity of judgements of adequacy, in the discussions of the
'soundness' of data, and of the 'reliability' and 'relevance' of in-
formation. An appreciation of what is involved in such judgements
may be gained from a consideration of a common and routine pro-
cedure in the formation of such judgements: statistical significance
tests. For statisticians do not simply say that a correlation is
'significant' or 'not significant'; rather, they will speak of sig-
nificance at a certain level. Those who have any craft skill in the
use of such tools will appreciate that the significance level to be
adopted is not assigned by God, but must be decided by the user.
The decision will be based on estimates of the direct costs and the
risks associated with each level. For each level of significance in-
volves the possibility of two sorts of error: of rejecting worthwhile
information, or of allowing dubious information to pass. The more
stringent the test chosen, the safer; but also the more costly, be-
cause of the extra time, care, and resources required for producing
material that will pass it. The choice of a particular level of sig-
nificance must depend on a judgement of what degree of safety is
required, for that component in its context in the total problem.
And this judgement must be based on general criteria of adequacy
applied to that particular situation. There can be no perfectly safe
test of the quality of the material, and neither can there be a cer-
tainly correct decision on the degree of stringency of the test.
In general, we can say that imposed criteria of adequacy are
necessary for scientific work because of the inconclusiveness of the
arguments used in science. This inconclusiveness follows from the
peculiar character of the objects of scientific inquiry: classes of
intellectually constructed things and events, the evidence for whose
properties is derived from particular experiences. To move from
the reported properties of particular samples, to the properties of
the classes they are intended to represent, a demonstration is
necessary; but a formally valid argument, yielding certainty or
truth, is impossible.
A scientific problem is thus incapable of having a solution which
is 'true'. Rather, the solution will be assessed for adequacy; and
for this every component must be so assessed. We can distinguish
two sorts of criteria and judgements of adequacy: those relating to
the argument, and those relating to the evidence. In the former
class are the tests of the various inferenceS which carry the argu-
ment through the statements of the properties of its objects to the
154 The Achievement ofScientific KnoT1Jledge
conclusions. Even a deductive argument is not exempt from testing
for adequacy; as soon as one passes beyond the simple syllogism in
sophistication, the obvious and intuitive tests fail to apply. Argu-
ments cast in mathematical form must satisfy criteria of 'rigour' to
the appropriate degree; and the rigour of physicists' mathematics
is not, and need not be, in the same class as that of the pure mathe-
maticians. Even in pure mathematics, disagreement over criteria of
adequacy have erupted at crucial points in the development of the
subject; the most.recent important case occurred at the beginning
of the present century, in connection with the use ofactually infinite
sets and constructions in mathematical proofs. 6 The criteria of
adequacy relating to evidence are more varied; for they control not
only the conditions of the production of data and information but
also the strength and fit of the evidence in its particular context. It
will frequendy be necessary for some of the evidence to be explained
6 At the end of the nineteenth century it could appear to mathematicians that they
had reached evolution's end, in the perfection of aiteria of adequacy. Thus Henri
Poin~: 'Today there remains in analysis only integers and finite or infinite systems
of integers, inter-related by a net of relations of equality or inequality. Mathematics
bas been aritbmetised.••• Have we at least attained absolute rigour? At each stage of
the evolution, our Cathers believed that they too had attained it. If they deceived them-
selves, do not we deceive ourselves as they did? ... We may say today that absolute
rigour bas been attained.' (Qpoted from E. T. Bell, Tile DevelojmInJI of M.tltnutits
(McGraw-HiD, New York and London, I94S), p. 34S.) Poincare was here reflecting on
the achievements of Weierstrass and his school in establishing the very tight patterns
of argument for 'potential-infinite' proofs in analysis. But the fa~de of perfection was
already revealing cracb in two places. First, one of the most basic objects of inquiry,
'real number, had been under investigation for some decades, and this 'foundations'
work was enQ)UJltering anomalies and oandoxes at an increasing rate. These were
related to the basic structure of logic, and to the 'actual infinite', which, after its com-
plete banishment in the middle of the century, was aeeping back into mathematical
arguments. At the tum of the century there was a 'foundations aisis', which led to the
inspired works of Russell, Hilbert, and Brouwer. The sequel to this was the aeation of
a new field of mathematics, in which many problems were set and solved, some of them
very deep (G6del's theorem being the outstanding example). But the basic problems
are still far from solution; see I. Lakatos, 'Infinite Regress and the Foundations of
Mathematics', AristotelitM Soeiet7 S.,plmtetJtMy VollIIM, 36 (1g62), ISS-8.f.. For a
personal acmunt of the reaction to the failure of all of the programmes to establish
'foundations', see John von Neumann, 'The Mathematician', Colletted Woris, vol. i
(Pergamon Press, 1961). He remarb that after the failure had been perceived by working
mathematicians, they 'decided to use the system anyway', for 'it stood on at least as
sound a foundation as, for example, the existence of the electron' (pp. S-6). Thus
working mathematicians operate with respect to the 'foundations' of their subject much
IS engineers do with nineteenth-century rigorous analysis: their techniques are designed
to avoid the known pitfalls, and no new ones have been detected. Since it is indisputable
that the 'foundations' of mathematics are not established, the absence of any new
'foundations crises' in the past haIf-<entury might be considered as evidence of a lack
of deep amc:eptuaI development in that period.
 Methods ISS
and defended explicidy; and these subsidiary arguments must also
meet criteria of adequacy appropriate to their function. Thus the
complexity of a solved problem is matched by that of the set of
relevant criteria of adequacy; and that set will depend closely on
the field of inquiry. Hence it is impossible to produce an explicit
list of criteria of adequacy applying to a wide class of problems.
This component of 'method' is thus a craft knowledge, but no less
essential to successful scientific inquiry than the formal, explicit
knowledge deposited in the public record.
This brings us back to the paradox of adequacy with which this
discussion began. It would appear that the certainty and objectivity
ofscientific knowledge are dissolving in a set of intuitive judgements
based on. principles which frequendy cannot even be made explicit,
let alone defended or tested. The easiest escape from the paradox
is to show that it is a false one; that the situation is not at all one of
objective truth being based on subjective guesses. For we have the
historical knowledge that some fields ofscience do achieve objectivity
and near-certairtty in their results, while others do not. In all of
them, there is an absence of formally valid proofs, and a presence
of conttolling criteria of adequacy. The difference between them
does not lie in this logical aspect of their arguments and conclusions,
but in the particular circumstances of their development.
The criteria of adequacy associated with a field are, however,
directly relevant to the sttength it attains. Indeed, they perform
an essential function, and a field is matured and effective only when
they perform it well. There the deficiencies created by the lack of
formally valid patterns of argument for basing general conclusions
on particular reports are remedied for practical purposes by the
accepted criteria of adequacy. For the tests of adequacy imposed on
the components of a problem at each phase of its development serve
for the avoidance or the timely recognition of the known pitfalls
in the way. In a matured field, unknown pitfalls are relatively rare,
and so the conclusions of the arguments will be well-established.
Although they are still not immune from the eventual discovery of
error or inadequacy, they provide a firm basis for immediately
subsequent work in the field. Also, the social tests on the materials
of a research report, necessary for their becoming accepted
as 'fact', can be applied in a sttaightforward manner, and a
direct transition from one status to the other can be the normal
event.
156 The Achievement ofScientific Knowledge
C;;terUl ofAtlefJUII&Y and the Maturing ofa Discipline
The function of criteria of adequacy is thus to make 'facts' p0s-
sible. It is worthwhile to examine this more closely; we can do so
in terms of the maturity or otherwise of a field of inquiry. The
condition of ineffectiveness or severe immaturity in a field can be
recognized (by those who wish to do so) through several symptoms.
In such a field, controversies on results range indiscriminately and
inconclusively from criticism of raw data to abstract methodology.
(Whether such controversies are common, depends on the social
character of the field; roughly, whether it is nascent or moribund.)
In this we see an effect of the weakness of the accepted criteria of
adequacy; they are not strong enough to channel debates on to
well-defined problems, nor to provide an agreed foundation for the
debates. Another symptom is that 'facts' do not exist. There is no
cumulative development of accepted results, stable under testing
and repetition, and invariant under changes in the problems in
which they are used. Instead, the results of each school or tendency
die with the problem around which its work was organized.
These two symptoms are related through a third, the ubiquity of
pitfalls. The failure of facts to be achieved is an effect of undetected
pitfalls. The conclusions of nearly all problems are soon seen to
contain errors; and although this is no hindrance at all to further
research in the same line, their materials are then valueless as in-
formation to be used in other work. But to conttovert a result is
not the same as to identify that flawed component in the argument
which vitiated it. Such flaws are rarely errors in logic; for as we
have seen, the basic patterns of argument in science are not formally
valid logical ones. The flaws are truly pitfalls, false assessments of
the quality of evidence or insufficiendy strong inferences; and they
are not easily identifiable even in retrospect, for the pattern of any
argument in a real scientific problem is too complex to permit of
an easy dissection. The presence of such pitfalls is caused by in-
sufficiendy strong criteria of adequacy. For when a particular pitfall
is identified, it is signposted by a particular criterion of adequacy:
at this point the work must be done rigorously, in a particular way,
lest disaster result. Conversely, a weak set of criteria of adequacy
fails to guard against pitfalls, and yields a situation where conclusions
of problems are illusory, and facts cannot exist.
The judgements of adequacy thus perform the same function in
scientific inquiry as the tests for quality control in industrial manu-
 Methods 157
facture; and the criteria ofadequacy stand in relation to them as the
standards of acceptable quality set for the tests. The solution of at
scientific problem is like a complex and delicate manufactured
product, in that a single weak component can destroy the whole
work. In manufacture, as in science, perfection is impossible; and
the standards of quality must be determined in the light of the
experience of the costs and risks associated with each level. The
difference between manufacture and scientific inquiry, in this con-
text, lies in the uniqueness of each scientific problem; even its
components are not all uniform, easily tested objects. Hence the
standards of quality control for scientific work are not capable of
being set out in handbook form, nor the tests carried out by a rou-
tine. The criteria and judgements of adequacy are necessarily an
informal, largely tacit knowledge; and in mastering and applying
them the successful scientist is a sophisticated and highly skilled
craftsman.
In the work of bringing a field towards maturity, an important
part lies in the strengthening of the criteria of adequacy. This is
not all, of course; the development of new tools, and the creation
of an appropriate social environment, are equally important. Nor
can the strengthening of the criteria of adequacy be done in an
abstract, automatic fashion, as by the attempted imitation of a
successful field. Such a strategy can produce totally misdirected
criteria of adequacy, leaving a field in worse condition than it was
originally. For the relevant criteria of adequacy are, as we have seen,
intimately related to the characteristic pitfalls of the problems in-
vestigated in the field; and these will be very particular to it, in its
objects of inquiry, its sources of data, its tools, and its patterns of
argument. Moreover, each particular criterion of adequacy must,
if it is to be effective, carry with it the craft knowledge of making
judgements of whether it is satisfied in particular cases, as well as
the tools and techniques for preparing materials of a quality
sufficient to satisfy it.
Over recent centuries, the founders and leaders of ineffective
disciplines have believed and proclaimed that the application of the
correct 'method' would quickly and automatically bring their
studies to a state of maturity and effectiveness. The conception of
method which has generally prevailed has been a combination of
two elements: a simplified pattern of argument (either inductive or
hypothetico-deduetive, depending on the current fashion), and a
 IS8 The Achievement ofScientific Knowledge
commitment to simple, preferably quantifiable data. Leaders of
influential schools in the social sciences have sincerely believed that
real science is done by putting masses of quantitative data through
a statistical sausage-machine, and then observing the Laws which
emerge.7 From such caricatures of the process of scientific inquiry
are derived criteria of adequacy which enforce an apparently
rigorous procedure of research,8 but whose results can but rarely
escape vacuity. Such programmes of reduction and mathematiza-
tion base their claims on the undoubted successes of the physical
sciences since the seventeenth century; but they ignore the long
series of dismal failures in applying this approach to the sciences of
life, thought, and society. And the latter history, largely unwritten,
is precisely as old as the success story: 'l'homme machine' of
Descartes's chimerical physiology is an exact contemporary with
Galileo's successful mechanics. The principle that each field of
inquiry has a degree of precision of argument appropriate to its
objects was laid down by Aristode; and the great mathematiCian
Gauss observed that 'lack of mathematical culture is revealed
nowhere so conspicuously, as in meaningless precision in numerical
computations'.9 If we extend these principles to criteria of adequacy
, Thus, T. S. Kuhn, 'The Function of Measurement in Modem Physical Science',
Isis, 5z (1961), 161-93, starts: 'At the University of CUcago, the fl9lde of the Social
Science Research Building bean Lord Kelvin's famous dictum: "H you cannot measure,
your knowledge is meagre and unsatisfactory.'" Gnnting the importance of measure-
ment in physical science, he asserts, 'I feel equally convinced that our most prevalent
notions both about the functions of measurement and about the source of its special
efficacy are derived largely from myth' (p. 161). He argues this thesis at length; and
the reader will see the influence of this essay on my own ideas. It is reprinted in H.
Woolf (ed.), /lflMlifielllitm: • History of 1M Me"';", of MetlswtmnJI i" 1M NtllUf'tll
.tIIl S«W Snmees (Dobbs Merrill, Indianapolis and New York, 1961).
• Much of the classic study, O. Morgenstern, 0" 1M Aeewtuy ofEetnUJmie Observll-
l;tmS (2nd edn., Princeton University Press, 1963), can be interpreted as an analysis of
misdirected aiteria of adequacy. On the basis of data that is deeply flawed by inherent
inaccuracies, 'applied' economists aggregate naticmal statistics to a high degree of
pseudo-prec:ision, and 'theoretical' emnomists construct models in which these very
dubious numbers are manipuJated in sophisticated and sensitive mathematical pro-
cedures.
9 Qpoted from O. Morgenstern, 0" 1M Ae"''''7 of EetnUJmie ObSerDillions, p. 99.
The German original reads: 'Oer Mangel an mathematischer Bildung gibt sich durch
nithts so auffallend zu erkennen, wie durch masslose Scharfe im Zahlenrechnen.' My
translation differs slighdy from that given in the text. For Aristode, see Nie0ttUl&1Ie""
Elllies, tt. W. D. Ross (Oxford, 1915), th. 3, l094b IZ-28: 'Our discussion will be
adequate if it bas as much clearness as the subject-matter admits of, for precision is not
to be sought for alike in all discussion, any more than in all the products ofthe aafts....
In the same spirit, therefore, should each type of statement be received; for it is the
mark oC an educated man to look Cor precision in each class of thinp just so far as the
 159Methods
in general, we can appreciate the fallacy of naive imitation as the
road to success.
One sign of a field having achieved maturity is a certain under-
lying stability, which persists through all the rapid changes in
results, problems, and even objects of inquiry. This stability is
revealed by the absence of new pitfalls, except at the outermost
frontiers of research; and, related to this, a set of appropriate and
stable criteria of adequacy. When these have remained in the inter-
personal channel of communication for some time, they become part
of the basic unselfconscious craft knowledge of the field; and the
particular judgements of adequacy which are made do not seem to
depend on anything but common sense. In these conditions the
very existence of criteria of adequacy can be overlooked; and philo-
sophers of science, basing their analyses on the experience of just
such fields, can remain in ignorance of their existence and im-
portance. But, as we have seen, in less matured fields they cannot be
taken for granted; and they will be of central importance in several
of our later discussions of the activity of science.
Criteria of Value
The criteria of value, and the judgements based upon them, form
an interesting contrast to those of adequacy. In those we saw a close
relation with philosophical questions of the possibility and nature
of scientific knowledge; while here we shall find ourselves involved
in problems of the social activity of science. The difference arises
from the functions of the judgements: the one is basically that of
assessment of work done, either a whole problem or a component;
and the other is an important determinant of the direction of future
work. For although judgements ofvalue are also made on completed
work, their crucial role is in the choice ofproblems to be investigated.
Thus, while the strength of the existing achievements of a field
depends (as we have seen) on its criteria of adequacy, its health and
future prospects are intimately related to its criteria of value. Be-
cause the criteria of value are complex, and necessarily involve
predictions ofthe future, the achievement ofappropriate judgements
of value is a delicate social task, even in matured fields where the
criteria of adequacy are well-established. In this respect, the social
problems of scientific inquiry in a field require attention even when
nature of the subject admits; it is evidendy equally foolish to accept probable reasoDiDs
from a mathematician as to demand from a rhetorician scientific proof's.t
160 The Achievement ofScientific Knowledge
the philosophical problems of scientific knowledge have been given
a practical resolution.
The exclusion of problems of value from the traditional philo-
sophy of science has its roots in the ideology of modern natural
science as it was formed through many generations of sttuggle. The
earliest conflict was on the philosophical plane; the pioneers of the
new philosophy of the seventeenth century conceived a world of
nature devoid of human properties, and accordingly rejected all
explanations in terms of 'final causes', or purposes of things and
events. Although they usually invoked the Deity in one way or
another, to give meaning to the world of nature in general, their
explanations of particular phenomena were cast in terms of purely
efficient causes. As the new science became established, the sttuggle
for the autonomy of its collective goals required an ideology which
denied any direct relation between the purposes of the community
at large, and the functions of the products of its work. Spokesmen
of science would remind lay audiences of the benefits which
ultimately flowed from their work, but would warn them that the
work must be pursued strictly for its own sake, lest the free play of
creativity be stifled. Thus science, conceived as a body of factual
knowledge, was cut off from considerations of value in two ways:
its assertions were purely descriptive, with no normative element;
and the considerations of social value by which all other human
activities are assessed were declared irrelevant. 1o The deeper
problems in which values enter, such as those concerning ethics
and morality, could for a long time be neglected because of the
favourable social conditions inside science and its very slight
responsibility for the social effects of industrial and military pro-
duction.
Certainly the facts presented in a textbook of mathematical or
physical science bear no signs ofbeing conditioned by considerations
of value of any sort. Even here, the appearance is illusory; for the
selection and presentation ofthese facts has been strongly influenced
by judgements of value. When one passes to books in the biological
sciences or applied sciences, considerations of value enter through
10 The defensive position of those who wanted to argue for the presence of values in
human knowledge is indicated by the tide of a book by the great psychologist W.
Kahler, Tile PIMt of VIII. in " Worltl of FMts (Liveright, New York; Kepn Paul,
London, 1939). A review of the litenture on this problem will be found in J. Leach,
'Explanation and Value Neutra1ity', BritisIJ]tnJnuIlfor PllilosopAy ofSrimct, 19 (1g68),
93-108.
 Methods 161

the term 'function'; devices are assessed in terms of their per-

formance of a function, and things occurring naturally are even
explained by the excellence of their performance of a particular
function. This is still a long way from bringing human purposes
into scientific explanation; but it is a reminder that the claim of
'science' to be value-free is based on very partial evidence.
In all the attempts to free science from every consideration of
value, no one has extended the argument to the activity of research.
For such an argument would not even have an initial plausibility;
choices and decisions must be made at every stage of scientific
inquiry, and such actions are impossible without judgements of
value. The extension of the boundaries of the known into the UD-
known does not take place like the spreading of a wet spot on a
piece of blotting paper. Without a strategy and tactics, a field of
scientific inquiry has as little chance of success as an army in
battle which is simply told to 'advance' .11 The statement of a prob-
lem to be solved is analogous to an objective to be taken; and the
selection of some problems for investigation, necessarily to the
exclusion of others, must be governed by competent judgements
based on sound principles. The judgements which determine such
decisions are those of value, feasibility, and cost. It is easy to see
that all three aspects of a project must be taken into consideration;
producing a result of little value will be a wasteful use of resources;
an unfeasible project, however great the value of its anticipated
solution, can lead to disaster; and excessive cost on a particular
project will starve others equally deserving. Given the feasibility of
a project, the decision to invest in it, rather than in others, will
be based on a balancing of value against cost, in comparison to
other competing projects. We notice immediately a feature of these
judgements which makes the decision process in science extremely
delicate: the things being balanced, value and cost, are incommen-
surable. For (neglecting remote effects of the result to be achieved)
the cost is of the inputs, measured in human effort and money;
while the value is of the output, and is complex and basically not
quantifiable. Because of this incommensurability, the 'planning' of
science cannot be reduced to a mathematical exercise in the same
way as the planning of a commercial venture.
For reasons I shall discuss later, the majority of competent
II See James B. Conant, 0" UtUkrst.ruli"K Slimet, ch. 4: 'Certain Principles of the
Tactics and Strategy of Science'.
 162 The Achievement ofScientific Knowledge
scientific workers do not frequendy make the choices which involve
an explicit assessment of the value of problems, and their balancing
against costs. But judgements of value are commonplace in the
informal, conversational communication among scientists, and are
used in the assessment of colleagues and of whole fields. A man will
be considered as something of a bore if he restricts himself to
working, however well and industriously, on problems whose
value, in terms of significance for future work, is moderate. Simi-
larly, neighbouring fields or rival schools will be subjected to judge-
ments of value, in terms of their potential for leading to new
advances; we shall see that such judgements are a very important
part of the government of science. These judgements are based on
criteria of value which are accepted, usually implicitly, within a
field. Work which deals with the objects proper to a particular field,
but which is so novel as to be incapable of judgement by the exist-
ing criteria of value, will be harshly treated. The pioneer whose
work lies outside the common experience of the field must expect
not only technical incomprehension by his colleagues, but also a
failure on their part to appreciate the value ofhis endeavours. To the
extent that the initiation of research in a field requires the invest-
ment of funds, and hence social approval, then the field may find
itself becoming ingrown and eventually stagnant through the
entirely natural working of these social judgements of value against
deeply original projects.

Components ofValue
The criteria of value applied in judgements on scientific problems
and results do not relate directly to the Good; but they are not
thereby deprived of interest. 12 We can distinguish three independ-
ent components in these criteria, two relating to the functions of the
completed piece of work, and the other to the purposes of the
scientist in undertaking it. Of these 'objective' components, the
first is the 'internal' one: to what extent the solved problem will,
or does, advance knowledge of the objects of inquiry of the field,
either directly or through suggested descendant-problems; this is
the dominant criterion of value in 'pure science'; and we note that
it is the least amenable to precise estimation, to say nothing of
12 The classic paper on aiteria of value is that of Alvin M. Weinberg, 'Criteria of
Scientific Cloice', M;"""., I (1963), 159-71, and ibid., Z (1964), 3-14; reprinted in
Rt/leaitnu till Big S~t (Perpmon Press, 196'7), pp. 65-122.
 Methods 163
reduction to quantitative measurement. The criteria of 'internal'
value adopted for a field will depend both on the social experience
of what has been successful in the past, and also on the general
conception of where the field is, or should be, heading at the point
of its development. For example, there may be a 'grand problem',
inherited from the founders of the field, whose eventual solution
defines the ruling criteria of value. 13 Alternatively, if the field is in
a phase of consolidation after a significant breakthrough, then
systematic studies of masses of particular details of a certain sort
will have considerable value, whereas previously the objects of
inquiry would have been changing too rapidly for such work to be
worthwhile. Within such general considerations of strategy, the
tactics of scientific progress are so varied as to offer a great many
possible functions for the results of solved problems, correspond-
ing to the variety of problem-situations out of which problems are
born. A knowledge of these functions, and an assignment of value
to each of them, is part of the craft knowledge of each field; and it
would be pointless to attempt an exhaustive formal list of them
covering all of science. Any attempt to make one sort of function,
such as 'explanation', 'prediction', or 'unification under a general
law', to be the unique and defining basis for criteria of value, is
bound to lead to an unrealistic account of scientific inquiry.
Parallel to this is the 'external' component of value: the con-
tribution that the completed project makes to the solution of prob-
lems, or the accomplishment of tasks, outside the given field. There
is a considerable variety in these external domains: they may in-
clude other fields of science, or technology, or (in earlier ages)
philosophy, or even the social or ideological needs ofthe community
of science or its larger society. Hitherto I have implicitly restricted
my discussion to problems whose dominant value is 'internal'. I

13 The classic example of such a 'grand problem' is the development of mathematical

methods to account for the perturbations and slow variations in the orbits of the moon
and planets; the problem was bequeathed by Newton to his successors, and culminated
in Laplace's demonstration that the variations are cyclic. Newton thought that an
occasional intervention by God would be necessary to prevent the system degenerating;
and this is probably the reference in Laplace'S legendary remark, 'Je n'ai pas besoin de
cette hypothese.' See A. Pannekoek,.If History of AstrOflOmY (Allen & Unwin, London,
1961), m. 30, pp. 297-307.
The alternative phenomenon is • commonplace in contemporary science; the in-
dividual researchers who join in the exploitation of a successful new technique may be
partly motivated by the following offashion; but in • healthy field the fashion is justified
by the need for detailed studies.
 164 The Achie'Demmt ofScientific Kno1lJledge
shall discuss the other sort in some detail later on; for the present,
it will be sufficient to mention that the nature of the dominant com-
ponent of value will strongly influence the criteria of adequacy
which are appropriate for the solution of the problem, as well as
the general path of evolution of the problem as it is conceived and
investigated.
In the ideology of science which became dominant towards the
end of the last century, any form of extemal component of value
was considered as a falling away from genuine standards of scien-
tific integrity. The corresponding ideal of 'pure science' was
certainly important in the effective establishment of refined norms
of behaviour; but as we face the problem of the intimate mixture of
extemal components of value in the work of industrialized science,
it is useful to see that completely 'pure' science is itself a recent
creation. First, we can say in general that internal components of
value can be appropriately applied only to the extent that there is
something to which the results of the inquiry can be internal.
When a field is still in the process of being created, purely technical
studies need to be supplemented (in varying degrees) by philo-
sophical inquiries on the one hand, and by the solution of practical
problems of organization and support on the other. For when a
field is too immature to yield facts from its investigations, nothing
is more futile than the attempt to amass 'positive' knowledge in the
absence of any reflection on the nature of the objects of that knowl-
edge or on the methods of its achievement. Also, even the first
genuine achievements can be jeopardized by the neglect of the tasks
of social construction in the field. These will include justifying its
existence to a lay audience; and the work done will inevitably, and
correctly, be conditioned by its function in this respect.
Until the nineteenth century, no field of natural science had such
intemal and social strength that its leaders could, or even wished
to, reject all extemal components of value from its work. The
'purity' of science seems to have been developed first in the uni-
versity environment of nineteenth-century Germany, where the
natural sciences struggled to win a place alongside the established
humanistic and philological disciplines. With the eventual establish-
ment of university science on a large scale in all advanced countries
in the earlier part of this century, the ideology of purity also
became a convenient means of preserving the autonomy of an
increasingly expensive social activity. By the entirely natural
 Methods 165
processes of the formation of a folk-history, the tradition of the
purity of science was extended back in time to the origins of modem
science, and across to all fields of scientific inquiry.
Scientific activity with only internal components of value is
certainly far simpler to manage and maintain than one where ex-
ternal components are effective. For in the distinction between
these two components, we have the root of the inevitable tension
between a scientific community and the lay society which supports
it. Speaking very roughly, we may say that the 'internal' components
of value are generally esoteric, and incomprehensible to the lay
public which pays the bills; while the completely 'external' com-
ponents of value in the present age are vulgar, and if dominant, can
eventuaIly debase and destroy the activity itsel~ In practice, the
health of a scientific community is maintained by a delicate and
inherently unstable compromise between the two.
An analogous compromise, effected through different institu-
tional structures, is necessary between these objective components
of value, and those relating to the purposes of the individual
scientist in undertaking a particular project, the 'personal' com-
ponents. It is quite conceivable for a man to pursue science with a
single-minded devotion, totally oblivious to the vulgar rewards of
power, prestige, and material comfort; and doubdess some do.
But such saints are usually a minority, even among the most gifted
scientists; and even such a person cannot totally exclude the personal
components of value from his judgements of problems. For the
most effective advancement of the field requires more than a single,
isolated worker; and winning support requires propaganda, both
to explain the work and to convince others of its value. To ignore
these social aspects of the task when presenting one's results, or
even when planning the strategy of a long campaign of investiga-
tion, is in effect to abdicate responsibility from the work, and to
be concerned only with one's private pleasure.
The extreme case of the inclusion of personal purposes in the
valuation of a problem is thus one where the motivation is quite
selfless. From this one can pass to a variety ofother sorts of personal
components of value, all of them justifiable in terms of acceptable
ethical principles. A university teacher may 'keep up' with some
research simply as a form of mental gymnastics, to prevent his
going stale and out of date in his field; in this case the cost of the
research is justified 'externally' by the maintenance of the quality
166 The Achievement ofScientific Kno1lJledge
of his teaching. The problems given to research students will
frequently be of very limited value by any objective standard; but
their prime function is to train the student in research methods, and
the feasibility ofa project for the student may well be a more critical
feature of it than its objective value. At the far extreme are scientists
whose criteria of value are nearly entirely personal; there we can
find men who follow, or who tty to create, successive fashions in
science for the enhancement of their careers, with the same facility
as that with which many women adjust the hemline of their skirts.
To demand that scientists abstain from including any personal
components in their valuation of problems would be ridiculous
and, to the extent that efforts were made to enforce such an impossible
standard of behaviour, they would only breed hypocrisy and cor-
ruption. Hence for proper valuations to be made, there must be
informal mechanisms for harmonizing objective functions and
personal purposes in science. These must be analogous to the
'hidden hand' of classical economic theory, which ensured that the
aggregate of private decisions, each set by enlightened self-interest,
yielded (in theory) the maximum social benefit. But for scientific
problems there is no simple, quantitative, automatically established
measure of value analogous to the market price of commodities.
Hence the mechanism for the achievement of proper valuations in
science must be more complex and subtle, and will necessarily be
more vulnerable to distortion.

Prohlems ofValue in the Social Activity ofScience

The argument as I have developed it so far might appear to be
leading to the same sort of paradoxes as those of adequacy. In this
case, they may be even more unpleasant. For there have been many
traditions in the philosophy of science which have weakened or
abandoned the claim to truth in natural science; while the long-
standing dichotomy between objective scientific facts and subjec-
tive value-judgements has hardly been challenged at all. Even though
I am not directly attacking that philosophical position here, I am
clearly arguing that complex judgements of value which are in
principle highly fallible, condition or even determine the selection
of those facts that aetuaIly come to be. To protect the world of
scientific fact from this invasion of judgements of value, one might
hope for an application of 'scientific method' itself, either made
unselfconsciously already, or deliberately through a 'science of
 Methods 167
science' in the future. Such a hope is a vain one, and a policy based
on it would be disastrous. My discussion of 'methods' up to now
has shown that there is no simple and automatic recipe for the
production of factual scientific knowledge; and the difficulties and
pitfalls of decision-making in the accomplishment of practical
tasks are, if anything, more severe. This is not to deny the import-
ance of social investigations into the structure of decision-making
in scientific communities; indeed, the present work may serve to
enhance the effectiveness of such studies. But an attempt to elim-
inate the elements of craft experience and personal wisdom in these
judgements and ·substitute for them a bureaucratic routine would
soon produce gross errors of planning from which committed
scientists would need to protect their field by clandestine research.
For the unknown is full of surprises; and to the extent that previous
experience is codified and made a rigid base for criteria of value in
decision-making, the further penetration into the unknown will be
blunted, and eventually reduced to the routine and the trivial.
The importance of the judgements of value, and the particularly
complex, delicate, and ever-changing character of the criteria of
value on which they are based, raise severe practical problems for
the government of science. As I showed earlier, the industrializa-
tion of science has produced a situation where decisions must be
taken more formally, and more centrally, than in earlier periods:
there must be a decision-machinery for channelling the capital in-
vestment in the production of scientific results, as a prerequisite to
the work itself. Hence the judgements of value which are influential
in setting the direction of new work are not merely the aggregate of
those of workers in the field; an extra, and very weighty component
consists of the judgements ofthose closest to the focuses of decision-
making power. To do their work well, such men must have unusual
wisdom and breadth of vision. They must not merely have the
imagination to make judgements of value on problems in fields
outside their special competence; but they must judge the value of
whole fields. And since they stand between the scientists and the
external agencies which are investing the funds, then the external
value of fields and of problems within them must be an important
element in their assessments.
Even during the 'acadomic' period of science, there were prob-
lems of government; but the 'aristocratic' principle of leadership
by distinguished men was usually prevented from enforcing an
 168 The Achievement ofScientific Kno1lJledge
unhealthy rigidity by the ability of the next generation to choose its
own patrons and problems. The execution of this choice depended
on the social character of the community; in the German culture
area, one could freely move between universities; and in Great
Britain a large proportion of the significant work was done by ama-
teurs. Only in France were the sciences organized in a centralized,
hierarchical, State-subsidized community; and it is significant that
after the single generation of the pupils of the great founders of the
French academic system, in the physical sciences at least the ~ne
was barren. 14
It would be a mistake to imagine the acts of value judgement to
be a constant and dominating feature of the work of science.
Scientists do not generally live in an existentialist agony of balancing
imponderable components of value of problems. Indeed, most
scientific workers need make few explicit assessments of value when
planning their work. But this does not mean that the choice of
problems is random or arbitrary. Rather, the majority are content
to accept the criteria of value which are currently dominant in their
community. But this makes the problems of government in science
more, rather than less acute. For the crucial tasks of determining
the directions of advance are concentrated among the aristocrats of
the community; these have experience and wisdom, but they may
also have a personal investment in work started in earlier decades,
and the burden of their political roles may leave them with in-
sufficient time for an informal, sensitive acquaintance with new
developments. Hitherto the community of science has functioned
14 The best synthetic account comparing the conditions and styles of scientific in-
quiry in Britain, France and Germany in the early nineteenth century is J. T. Merz,
A History of £.011." TAougAt i" tile Ni.te""A C"".y, 4 vols (1904-12; reprinted
Dover Publications, New York, 1965), voL i, ch. 1-3, pp. 89-301, especially summary,
pp. 298-301. The nineteenth-century historian Buckle provides a good example of the
traditional English distrust of centralization in cultun1 matters. Proceeding from his
critical analysis of the state of intellectual culture under Louis XIV of France (where,
unlike all previous historians, he showed that it was abysmal), Buckle concluded that
the 'protective spirit' of the State applied to science through oflicial patronage will
inevitably be destructive in its effects. He posed the problem of decision: if merit is
decided by the State, it will be done incompetendy; but if it is decided by the leaders
of the community receiving the rewards, it will be 'disgraceful' and degrading. Citing
the earlier cases of the reigns of Augustus and Pope Leo X, he said 'in each of these ages
there was much apparent splendour, immediately succeeded by sudden ruin'. The editor
of the later edition of the work, the freethinker John M. Robertson, disagreed violendy
with Buckle's whole argument. See Henry Thomas Buckle, l""otlwt;tm to tile History
ofCiviliSfllitm i" E",lfIrul (1857-1861), new and revised edition (Routledge, LondoD;
Dutton, New York, no date). I am indebted to R. G. A. Dolby for this reference.
 Methods 16g
well without explicit institutions for the exercise of influence from
below; but whether such informality in this respect is sti1l appro-
priate when so much else is institutionalized is a question worth
considering.
Generalizations of· Method
A full discussion of 'methods' requires mention of two more
senses of the term, which are important in the historical development
of modem science, and which are relevant whenever a new field is
brought into being. At crucial points in the development of science,
there will frequently appear proclamations of 'method', including
in their scope not only the sorts of methods we· have already dis-
cussed, but the objects of inquiry as well. Thus they serve to give
a complete definition of a field of inquiry, or of a whole discipline.
The classic statements on the new philosophy of nature by Galileo,
Descartes, and Bacon fall into this class; and the major works of
the latter two were explicitly treatises on 'method'. Such proclama-
tions form an exception to the general rule that the principles and
precepts of method are transmitted through the interpersonal
channel of communication. The reasons for this lie in the excep-
tional character of their audience and their message. For they are
not intended primarily (if at all) for pupils in a field, serving to
supplement the craft knowledge they have received informally.
Rather, their audience is a wider one, including a part of the general
public; and their function is to explain and justify their subject to
this wider audience, and to draw support and recruits from it.
Similarly, the message will have a minimum of technical detail and
precepts for solving particular problems; rather there will be general
principles, supposedly applicable to a wide (or even universal) class
of problems. IS
The special character of these proclamations of method also
15 Although the three 'essays' accompanying Descartes'. DiS&fJWS 4e 111 MnluHll are
on mathematics and natunl philosophy, Descartes'. programme was much more ex-
tensive. The first tide Cor the DiS&OfII's itself was Le Projn 4'. . seime, "";TJII'SlIIe pi
puisse ,lever notr, fill'.'II StnI pllU lUll" tle".1 tle perftairm. See It. Descartes, DiS&fJWl
tle 111 M,,1uNle, p. m. The Utopian and miJlennarian aspects of Francis Bacon'. thought
are fairly well known, as well as his concern for mannen and monls. Although his
NtnJIIIII Or,II""", is primarily concerned with uatunl philoeophy, Bacon elsewhere
expressed his belief that the 'primary uiools' of philoeophy apply equally to an aspects
of God'. creation, including the social and montu wellu the uatunl. See the section
on 'Philosophia prima' in the De Allplllllil S~ Book III, cit. I; Woril, iv,
336-40 (translation).
 170 The Achievement ofScientific Knowledge
derives from their particular concerns. The technical tasks of in-
vestigating problems in a field are subsidiary to a larger philo-
sophical mission; and this may be imbued with a touch of the
prophetic as well. The experience on which these proclamations
are based will correspond to these deeper concerns: not so much
doing a well-defined job and then reflecting on the way it was done,
as attempting to create a new sort of work altogether, and mastering
its basic principles along with its practice. The investment ·of one's
life and talents in such hazardous work demands a deep commit-
ment to the success of its results, both in the achievement of new
knowledge and in the diffusion of the new approach. Hence (unless
the knowledge is esoteric, a rarity in our present civilization and
automatically excluded from our natural science), there will be a
natural tendency for the author to believe that the 'method' charac-
terizing the new approach is as easily communicated to a wider
audience as are the scientific results of the work. The subtle, par-
ticular, tacit component of the work will accordingly be neglected
in the reflection on its principles, since this would obscure the
message and decrease its prophetic effect. Because of this, the
proclamations on method have strengths and limitations opposite
to the accounts of the 'art of inquiry' in particular fields. For the
latter sort is strongest on precepts of craft practice and tends to
confusion or vacuity when the author philosophizes; while the great
proclamations provide general illumination and philosophical depth,
but their explicit prescriptions on method are sterile or banal when
applied to the craft work of investigating scientific problems. 16
Such general proclamations of method are usually intended as
clear calls to essentially straightforward action; but when they are
received by an audience, they are rightly interpreted as philosophical
theses. For their nature demands that they include an analysis of
the knowledge of their domain of inquiry, and also that they have
clarity and coherence as well as depth of insight. Hence they are
judged by the criteria of adequacy and value prevailing in the rele-
vant branch of philosophy. Their authors may well be dismayed by
16 The 'four ndes' of method given by Descartes in his Distows tle 111 Methotle
(Second Part) were probably a fair sketch of the personal style of problem..-olving he
had developed IS a precocious child; but IS announced they were of little use to anyone
else engaging on detailed philosophical or experimental work. By contrast, Bacon at-
tempted, in the Second Book of the NtnJIIIII o,KII",., (ttanslation in Woris, iv, 11g-263),
to provide a complete list of the types of judgements involved in the various stages of
inquiry in his style; but here the edifice of distinctions coUapsed under its own weight.
 Methods 171

such a reception, especially since there is a tendency among deeply

original philosophers to see their work as finishing off that un-
satisfactory discipline once and for all, and laying the foundations
for solid scientific work in its place. But philosophy always refuses
to lie down and die, and the classic proclamations of method take
on a philosophical career of their own, producing a crop of descend-
ant-problems within the ancient field of methodology. 17
As a field of philosophy, this is a most difficult one in which to
maintain traditions of work of high quality. For the criteria of
adequacy are necessarily of two conflicting sorts; both a philo-
sophically coherent analysis, and a perceptive insight into the actual
practice of science, are required. The criteria of value for problems
in methodology are correspondingly twofold. If there is no con-
tinuing delicate balance between these two aspects of the inquiry,
or even worse if there is no awareness of the distinction between
them, discussions of methodology will be largely sterile in practice,
or conceptually muddled, or both. Over recent years, most dis-
cussions of 'scientific method' have suffered from a confusion over
the methods by which they themselves are governed. In particular,
the relation between their discourse, and the practice of science, is
but rarely made explicit, and the ambiguities in the assertion, 'The
scientist does ...' are never clarified. 18 For this might refer to what
scientists actuaIly do and know that they do, most of the time; or
it might refer to principles which are implicit in their work (as
syllogistic reasoning in ordinary argument); or again it might refer
only to a 'rational reconstruction' of their activity, for revealing the
logical structure of scientific results. Further, the assertion might
be prescriptive rather than descriptive, and would thus indicate
what scientists should do in one of the above senses. Evidence for
the presence of these confusions is provided by some interpretations
17 For histories of this &c1d of inquiry, see Ie. McKeon, 'Philosophy and Scientific
Methods',JounuU oftile History of/MIlS, 27 (1g66), 1-22; and also L. Laudan, 'Theories
of Scientific Method from Plato to Mach', History of SGinlee, 7 (1g68), 1-65.
II For strict 'logical empiricism', the question of what it is talking about, has been
effectively probed by Stephen Toulmin, in 'Are the Principles of Logical Empiricism
Relevant to the Actual Work of Science?', StinJtiji& Amerit"", 214 (February 1966),
129-33, a review of C. G. Hempel, Aspe,t, of StinJtiji& Ezplilutitm IIIIIl Otller ESIIIYI
i" tile PlJilolfJpIJy ofSlimee (The Free Press, Glencoe, 1965). See also the correspondence
between E. Nagel and Toulmin, in the Aprillg66 issue, pp. 8-1 I. The sharp dichotomy
between the.'coDtext of dischvery' and the 'context of justification' is DOW coming under
auaek even by philosophen ofscience raised in the tradition in which it was established.
See P. Caws, 'The Structure of Discovery', SGinlee, 166 (1969), 1375-80.
 172 The Achievement ofScientific Kno1lJledge
of Popper's philosophy of science; in these, it is asserted that scien-
tists normally do, and always should, make the refutation of an
initial hypothesis the goal of their work as in a research project. 19
Yet although the field of methodology is difficult, its problems
are not vacuous. As we have seen, even though scientific work may
appear to be nothing more than common sense to the majority of
its practitioners, it is a highly sophisticated activity dea1ing with
very artificial objects. At any point of genuine innovation (without
which science becomes stagnant and dies) the inherited unselfcon-
scious craft knowledge of objects and methods is inadequate to the
task. A deep innovation in science necessarily involves an innova-
tion in methods; and if the innovation in methods is itself deep, its
explanation and justification will involve arguments in method-
ology.
Taken all together, the 'methods' of a field of enquiry govern its
work, from the most particular techniques of tool-using, through
the criteria by which the quality of work is assessed, to the most
general principles underlying the practice of the field. zo They thus
form a natural correlative to the objects of inquiry; the set of these
latter defines the subject-matter of the field, while the former define
the ways in which this subject-matter becomes a part of human
knowledge.
Without a body of methods of this sort, the social activity of
science could not exist. It is possible for a man of genius, working
single-handed and in isolation, to create a new field of study,
developing his methods intuitively or even unselfconsciously as he
defines and solves his new problems. But for a field to reach a
matured and effective state, there must exist in the interpersonal
.t In his TOIIHIrth ." HistoritJl"flllIy ofSeinI&e (The Hague, Mouton, 1963) (History
IIIIIl Tlwry, Beiheft 2), J. Apssi proposes that historians of science should proceed by
testing and refuting each other'. bold conjectures (p. 74); and his examples indicate
that pat scientific discomics are in fad: made by refutation of previously accepted
hypothesa
.. To cite a very simple example, in 'rational mechanics' the objects are those of
Euclidean pometry IS CIU'iched by the ameepts of velocity, acceleration, force, mass,
and energy, with otbcn depending on the special problems, and all of them linked
by deliDitioas~ The methods include in the fint place the accepted procedures for using
certain mathematical tools (historically, differential equations but now including other
parts of mathematics), where the procedures include accepted interpretations of the
mathematical symbols, and appropriate stmdards of mathematical rigour. Other
methods include the more subtle criteria of adequacy reprding the relation of the
results of an interpreted mathematical argument, with aperiment or experience. See
C. Truadc11, ESlllys ill ,lie HUt.y of Meelumi&s (Springer-Verlag, 1968), pp. 334-40.
 Methods 173
channel of communication a large body of knowledge of the
different sorts of methods; for men of genius are too rare, and too
individual in their style, to be sufficient for the production of that
collection of solid facts without which a field cannot steadily in-
crease and deepen knowledge. This phase of the development of a
field requires numerous craftsmen who are competent to operate
effectively with inherited skills. Of these, only a minority can have
the talent and commitment necessary for teaching themselves such
skills from their own successes and failures. Without a period
of training for research, where they can absorb an existing body
of methods, they would only blunder through ill-conceived and
misdirected projects, and disrupt the delicate system of social be-
haviour which enables a field to exist. But with an informal instruc-
tion in methods, they can do their essential work. In this way, an
established body of methods has a function analogous to one of
those of tools and their associated techniques. They make possible
a downgrading of tasks, so that there can be a division of labour,
and the possibility of that complex social organization which is
necessary for the effective advancement of knowledge on a broad
front.

Methods liS Crllft Kno1lJledge

Our study of 'methods' has shown that they are a very hetero-
geneous collection of things, governing different aspects of the
complex social activity of scientific inquiry. But they have one
feature in common: except for straightforward techniques, they are
all largely informal or even tacit knowledge; and they are trans-
mitted through the interpersonal channel of communication, rather
than through the public channel of printed reports. In these re-
spects they form a conttast to the objects of inquiry of a field; and
a study of the full set of contrasts and relations between these two
opposite components will enhance our understanding both of the
activity of science and of its products.
I have already argued that the objects of scientific inquiry are
classes of intellectually constructed things and events. They must
be mainly public and explicit, rather than informal and tacit, if they
are to perform their function in the achievement of this particular
sort of knowledge. For scientific knowledge, as it has been under-
stood since Aristotle, seeks a rational understanding of its objects
 174 The Achievement ofScientific KnoJl)ledge
rather than a dialogue or communion between them and the
knower. Even if the objects are human, or are conceived as endowed
with some human properties, the contact between knower and
object is not an empathetic one of shared inner experience, but an
indirect one, achieved through the conclusion of an argument.
Although scientific knowledge is ultimately based on particular
experiences of interactions with an external world, any report of a
single such experience is only a fragment, nearly meaningless and
useless in itself. For the significance of such a report is not what
happened to the personal agent, but the clues it can give to the
constitution and workings of the external world. Similarly, research
reports are not designed for the communication of the personal
experience of the agent, but to provide materials for others working
in a similar fashion. Hence the objects of discourse must be so
explicit that there is a sufficient overlap among the private under-
standings ofthem for them to be taken over, used, and adapted, with
success. In this sort of work, ambiguities and a variety of interpre-
tations are not a source of richness and depth, but of waste and
error.
The line of demarcation between 'science' in this extended sense,
and other sorts of knowledge, is not between the human and non-
human worlds as domains of inquiry. For until a few centuries ago,
the dominant traditions in the study of the natural world saw it as
human in character and suffused with divinity; the alchemists and
magicians tried to establish a dialogue or communion with it. And
in the disciplines studying man and his creations, including psy-
chology, history, and even theology and literary criticism, are
'sciences' in this sense. Any sort of scholar argues from evidence to
a conclusion; and even if his purpose is to help his audience achieve
a sympathetic understanding of a personal experience, his particular
task is accomplished with an argument rather than an in-
cantation.
The objects of scientific inquiry cannot be perfectly public and
explicit; and we should review the limits and borderline cases to
their public character. For example, I remarked earlier that the
first achievement of a deep insight on a scientific problem can be an
intensely personal experience, sometimes nearly mystical in its
quality. Also, the objects of scientific knowledge cannot be taught
purely by demonstration; especially in elementary instruction, they
are conveyed by indocttination and manipulation, and become part
 Methods 175
of the unreflective craft knowledge of the trained person. Even in
advanced research, each scientist has his own private understanding
of his objects of inquiry, deriving from his unique craftsman's
experience of working with them. Finally, those objects which
become established as concepts of genuine scientific knowledge
never cease to undergo change, and they carry with them a history
of ambiguity and concealed contradiction.
Nor is the contrast to other sorts of knowledge an absolute one.
Even, for example, most poetry and the reports of mystical ex-
periences are intended as social possessions. Also they exist in a
language which inevitably is culturally conditioned, and in part
intellectuaIly constructed. But the knowledge which they yield is a
fabric of interwoven private experiences, and the special, unique
character of each of them, and the ambiguity in their messages, is
a conttibution to the richness of the whole. We may illuminate the
contrast between such 'personal' knowledge and 'scientific' know-
ledge by considering a report of an inspired insight into a scientific
problem. If this is well written, it may stir the imagination and the
emotions, and enhance the reader's appreciation of the workings of
the mind of the scientist. It may also yield evidence for ,studies in
the history of science or in the psychology of invention. But as
material for the further development of knowledge in that field of
science, the report is of only speculative value. For it is the content
ofthe insight, not its quality as an experience, that is relevant to this
function; and until the insight is proved to have a content which
can be understood and applied by others independently of any
sympathy with the author's experience, it remains in the realm of
poetry rather than science.
The sort of practical craft knowledge which is embodied in the
methods of scientific work is midway between the 'scientific' and the
'personal' knowledge which I have put into sharp contrast. For it is
incapable of explicit specification, and yet it is a social possession;
although the principles and precepts of method cannot be estab-
lished by an argument of the sort characteristic of scientific know-
ledge, there must be a uniformity of understanding of them, as
reflected in practice, fqr a field to exist. Methods "are midway in
another sense; for the very private experience of a working scientist
with his materials is converted into a public, social possession
through operations governed by the body of informal, interper-
sonally communicated methods.
 176 The Achievement ofScientific Kno1lJledge
The TJl)O Channels ofCommtmiGat;on
It is useful to compare the channels of communication whereby
these two quite different sorts of knowledge, the properties of the
objects of inquiry, and the methods, are transmitted. Research
reports, embodying new conclusions about the objects of inquiry,
go mainly through a 'public' 'Channel of papers published in recog-
nized journals. This provides an impersonal dissemination of
tested materials to an unresttieted audience; anyone with the
technical competence to understand a paper is entitled to read and
use it. Although the procedure ofsuch publication may seem simple,
the public channel aetually accomplishes a variety of tasks, each
serving a different social function. The most obvious work of the
public channel is the widespread and rapid dissemination of new
results; this is not only facilitates the advance of the field, but also
exposes the work of every individual and school to the critical
scrutiny of colleagues. Supplementary to this is the role of the
journals as an archive, a permanently open repository of the past
achievements of the field, where they are available for later use.
The system of publication in journals also serves to protect the
personal interests of individuals; for a published paper is the
property of its author, and the system of citations is designed to
ensure that his property is recognized. Finally, none of these
functions could be performed were it not for one other: that of
quality control. If there were not a test of each paper before its
acceptance by a journal, then every intending user would be forced
to examine it at length before investing any of his resources in work
which relied on it. Under such circumstances, the co-operative
work of science as we know it could not take place. Similarly, the
archive would consist of unauthenticated claims to results, and as
such would be virtually useless. Moreover, the intellectual property
of the author which is embodied in a published paper would be
nearly worthless, since the status of publication would carry with
it no indication that the result was genuine.
These various functions of the public channel are quite different
in character, and each relates to the private purposes of scientists
in its own way. Nor is it guaranteed that excellence in the perform-
ance of one such function will produce excellence in the others. We
shall later see how the public channel is sensitive and vulnerable,
and how the performance of its various functions, and the nature of
its contents, are influenced by the social practices and attitudes of
 177 Metlwtls
the community it serves. But provided that it works well, as it has
generally done in the past, its contents comprise both news and
archive, and are the materials from which genuine scientific knowl-
edge is eventually achieved and recognized.
For the body of methods ofa field, such a public channel, with its
controls, is simply not available. There can be a public transmission
of results in methodology; but as we have seen, this is not at all
the same as the craft knowledge which governs the work being done.
Also, criticisms of results and of methods will appear in the public
channel, partly in the body of the argument of research reports, and
partly in the 'correspondence' section of a few journals. But, even
granting the importance of such channels (especially for the dis-
cussion of the political and social problems facing contemporary
science),21 such materials are not research reports like the principal
contents of the public channel and they cannot achieve the 8u'Jdety
and directness which is possible within the interpersonal channel.
Hence, even with these exceptions, we can say that the transmission
of methods is accomplished almost entirely within the interpersonal
channel, requiring personal contact and a measure of personal
sympathy between the parties. What is transmitted will be partly
explicit, but partly tacit; principle, precept, and example are all
mixed together. There is no substitute for such personal com-
munication; messages whose transmission requires a prior formula-
tion and clarification of ideas (as even in a letter to a colleague),
will necessarily be impoverished in their content of private craft
knowledge. A period of intense interpersonal communication be-
tween members is necessary for the formation of a school; only in
this way can a full body of methods, including the personal style of
masters, be communicated in the context of frequent informal
discussions of work in progress. Gatherings of scientists perform a
similar function of interpersonal communication of methods,
although in a diluted form; but even a brief remark assessing some
recent work can convey significant information on developments in
the ruling criteria of adequacy and value.
The two channels of communication are not entirely separate in
21 The contrast in this respect between N."" and Slime" the leading weekly general
scientific journals in Britain and America repectivc1y, is iDstruetive. SeinI&, bu a lively
correspondence section at the front of the issue, while Nfl"', prints only a few letters,
commenting on specific points in artides, tucbd away at the back. The 'Icademic market
place' thus bu certain advantages over the 'club' for the public dilcussion of COIIUDOD
problems.
 1,8 The Achievement ofScientific KnoJl)ledge
their contents. The public account of a piece of work, embodied in
its research report, must include enough description of the tools
and techniques to enable a competent person to reproduce the
argument and re-create the evidence, even if he could not have
conceived the original problem on his own. For without the explicit
communication of this component of method, the materials of the
report would not be capable of becoming facts. Conversely, the
gap between the private investigation of a problem and the public
description of its solution is frequently bridged by a communica-
tion in the interpersonal channel: verbal or even written descrip-
tions of the work for colleagues at all levels of formality below that
of submission to a journal for refereeing and publication. Some
measure of such interpersonal communication of achieved results is
necessary for efficient co-operative work in a field; but an impor-
tant part of the 'information crisis' in contemporary industrialized
science is the substitution of this hybrid channel for the trans-
mission of cumulative series of results at the expense of the public
channel. The absence of controls in this channel can lead to a de-
terioration of quality, as I have discussed already. Similarly, the
public certification of the intellectual property embodied in a
research report is not achieved unambiguously and automatically;
and if results become public in the interpersonal channel only, the
social task of the protection of this property is deprived of its
institutional support, and depends more heavily on the integrity
and skill of all the members of the community.
The relation of the two channels of transmission is even more
intimate than the overlapping of contents would indicate. For the
materials which appear in the public channel are strongly con-
ditioned, in their selection, by judgements made in accordance with
principles transmitted in the private channel. To enter the public
channel, a research report must be certified by a referee as passing,
in respect of both adequacy and value. Yet the criteria on whose
basis editors and referees assess papers appear nowhere in the public
channel itself. They are a part, a most sophisticated and subtle
part, of the craft knowledge which constitutes the methods of a field.
If they are inappropriate for the development of a field, the material
which appears in its public channel will be correspondingly dis-
torted.
On its side, the interpersonal channel of communication is not
independent of the public channel; but the two patterns ofinfluence
 Methods 179
are not symmetrical. Ifthe public channel, considered as the archive
of results, shows scanty progress in a particular field, then aitics
may call into question some part of the body of methods governing
the work in the field. But such criticisms of methods cannot be
derived as conclusions to arguments resting on conttolled experience;
if nothing else, the assessment of the quality of progress is itself a
judgement based on criteria of adequacy and value. Hence the
mechanisms for the conttol of methods, and of their interpersonal
channel of communication, use materials characteristic of that
channel itself, and not those of the public channel. This holds eveD
when some criticism and debate is formally published; for it ap-
pears as such, and not as research reports. Thus the results which
appear in the public channel are passive. They can be used for a
variety of functions, including the provision of evidence for the
criticism of themselves. But by the nature of its contents, the public
channel cannot directly 'influence' anything; and the relation be-
tween the two channels is not one of automatic mutual controls.
This asymmetry in the relations between the two channels of
communication has important consequences for the government of
science. For it is possible for a field to be diseased; for the ruling
methods to be such that ineffective or shoddy work is accepted or
even enforced. But the establishment of the existence of such a
condition, and even more the attempts to reform it, are problems of
a different sort from those which the field is devoted to investigating.
Moreover, while gross errors in technique can be detected by the
discovery of pitfalls in published work, the more subde components
of method are not easily accessible to anyone who has not already
invested a part of his life in working in the field. Thus, reforming a
diseased field, or arresting the incipient decline of a healthy one, is
a task of great delicacy. It requires a sense of integrity, and a com-
mitment to good work, among a significant section of the members
of the field; and committed leaders with scientific ability and
political skill. No quantity of published research reports, nor even
an apparatus of institutional structures, can do anything to main-
tain or restore the health of a field in the absence of this essential
ethical element operating through the interpersonal channel of
communication.
In conclusion, we may consider the two channels of communica-
tion and their contents as a pair of interpenettating opposites. The
one disttibutes and preserves the results of the work, while the
 ISo The Achievement ofScientijie K,,01lJledge
other governs the work itself; one is public and explicit, while the
other is informal and interpersonal. The contents of the public
channel are in principle permanent, and exist independendy of the
circumstan~ or ultimate fate of the work which produced them;
while the body of methods, bound to very particular personal
experience (both technical and social) directly control the future
contents of the public channel. The results of scientific inquiry are
in principle based on controlled experience and rigorous argument;
but the methods governing the inquiry itselfare a particularly; subtle
craft knowledge, different in nature from scientific knowledge.
Thus we have elaborated the apparent paradox that scientific
knowledge is achieved through an activity based on knowledge
which is not scientific in character. If we restrict our attention to
the investigation of particular problems, the paradox has no
resolution; but I will later show how the complex social processes
whereby the contents of research reports enter a new phase of de-
velopment can yield a practical resolution of this paradox along
with the others.
 6
FACTS AND THEIR EVOLUTION

ONCE we have achieved the slightest historical perspective on

science, and cease to view it as the steady and certain accumulation
of increasing heaps of true facts about the natural world, a number
of paradoxes present themselves to us. First, there is the conttast
between the highly personal endeavour required for a penetrating
inquiry into the natural world, always fallible and governed by
mainly tacit craft methods; and the public, objective, impersonal
knowledge which eventually issues from that work. Similar to this
conttast is that between the ephemeral structures in which scientific
inquiry is undertaken, problems succeeding each other with great
rapidity and theories and concepts being ploughed under at a
perceptible rate; against the steadily increasing depth and power of
the knowledge which ~mehow remains and grows amidst the swirl-
ing currents of the day-to-day work. And, finally, we have already
observed the impossibility of 'proving' a single result in science,
or even of establishing its 'probability' of being true; and yet the
stock of permanent knowledge, absorbing all reinterpretations and
surviving all attempts at refutation, survives and grows.
There is no magic formula to explain the success of natural
science; neither some special property ofnatural things which makes
them uniquely accessible to human reason; nor a victorious 'method'
whose application has laid bare the secrets of the natural world, and
which can now be turned to other fields at will. Rather, the scientific
knowledge we possess is the result ofa social endeavour, which over
the centuries has developed an approach appropriate to its limited
goals, and where the work of each individual is informed and con-
trolled by that of his colleagues in this endeavour, of the past, the
present and the future. In the previous chapters I discussed the
phase of the achievement of knowledge which is largely the work of
7-&u.P.
182 The Achievement ofScientific K,,01lJledge
an individual or a small group: the craftsman's work of investi-
gating problems. The resolution of the paradoxes in the character of
achieved scientific knowledge will come from an analysis of the
further development of a solved scientific problem, when it emerges
from the workshop of its creator, and takes on a life of its own in the
community of science.
In this chapter I shall first discuss the processes and tests whereby
acomponent ofa scientific result comes to be recognized and accepted
as a 'fact'; from this analysis we may achieve a clearer understanding
of that basic but difficult concept. I Not all facts are, or become,
genuine scientific knowledge; they must survive lengthy and rigorous
processes of testing and transformation. These take place in the
course of the evolution of the different components of a solved
problem. From a comparison of the different ways in which the
different components evolve, we will be able to identify an essential
feature of those facts which survive to become scientific knowledge.

The Research Report and its Tests

The achievement of genuine scientific knowledge is the ultimate
goal of the activity of scientific enquiry; but does not come quickly
or easily. The completion of the work of investigating a scientific
problem, even when the criteria of value and adequacy are satisfied,
does not necessarily yield knowledge. Indeed, it does not even neces-
sarily yield facts. For the product of a completed scientific investi-
gation, we should best use a term as neutral as 'research report'. As
it stands, the report is literally not to be trusted; and in practice it
is not. So great are the possibilities oferror, especially in work which
is so complex and subtle, and whose acceptable outcome is so
important to the individual who performs it, that the research re-
port must be checked by an impartial assessor before it is released to
the community. This is commonly done by a referee for a scientific
journal; only after the referee has certified the problem as being of
sufficient value, and the work as having been performed to the ruling
I I should make it clear at the outset that the term 'fact' IS I use it refen to assertions of
a particular 1Ort, and fIOl to a state of affairs in the extema1 world. Much of the recent
philosophy of science is governed by an implicit usumptiOD that there is generally a
perfect aHlespondence between non-trivial assertions about the world, and their real
objectB; 10 that 'fact' is taken IS referring primarily to the world and then to its des-
aiption. For me an important problem is how such a correspondence. even an imperfect
one, ever coma to be achieved and also remgnized; and 80 I must use the term 'fact'
with peater precision.
 Facts and their Evolution 183
criteria of adequacy, can it appear in the literature. A research report
which is distributed without this certificate is, or traditionally has
been, considered as nearly worthless. I have already discussed the
significance of the partial breakdown of the system of certification,
as a symptom of the new social conditions of industrialized science,
and as a danger in itself.
Refereeing a paper is a special skill of a matured scientist. It is
hardly ever possible to reproduce the data on which the evidence is
based, and it is frequently onerous to follow a complex argument
through all its fine structure. So the referee must use his personal
knowledge of the craft to form a judgement on the adequacy of the
work, using as evidence the reports and descriptions in the text
before him. In this work the scientist develops skills which are
much closer to those of an historian than is generally realized. Also,
the filter applied by the refereeing system is inevitably conservative
in its effects. Since really pioneering work involves a recasting of
criteria ofadequacy and value, the referee will generally find it below
standard in some respect in the accepted system. Since the style of
such work is radically different from that of straightforward medi-
ocrity, it will be at risk of assessment as belonging to a much more
perjorative category: that of the crank.
It might be thought that a research report, duly certified and
published in a recognized journal, can be accepted as embodying a
unit of scientific knowledge. Indeed, the belief in such a rapid and
straightforward process of the achievement of scientific knowledge
.was dominant until recently, and is still widely held. But my analysis
of the referee's work indicates at least that great things are not won
so cheaply. For all the referee can assure is that in his judgement the
problem is of value and the work adequately performed. The true
value of the problem can be determined only through the further
development of the field; and the real adequacy of the work can be
properly assessed only by the more demanding tests ofrepetition and
application.
Hence there is another phase of the testing of a research report,
performed in the natural course of scientific work, outside of any
institutional framework. The test of value is the most crucial; and
it will generally be applied more rigorously by the community of
scientists than by an individual referee. Fot the rejection of a paper
by a referee on grounds of insufficient interest is a positive aet,
requiring a justification. But for a scientist to include a particular
 184 The Achievement ofScimtijie Kno1lJledge
paper in the overwhelmingly large class ofthose which do not interest
him, requires no apology nor indeed much cogitation. Almost every
paper is uninteresting to almost all those who happen to see it; and
the automatic, communal test of value yields a negative result when
a particular paper happens to be of interest to absolutely no one.
As a candidate for the status ofscientific knowledge, it has then been
killed. It may continue its career in a political function, as an entry
in the scientist's list of publications, but as a contribution to the
advancement ofknowledge it has died, precisely because it is unwept,
unhonoured, and unsung.
For a research report that passes the communal test ofvalue, there
is another test of adequacy, again more rigorously applied than by
the referee. This is in its use. Here, the judgement is not whether it
seems all right, but whether the results can be reproduced and built
upon. Ifthe test is passed the research report continues in being, as a
citation in later papers, so long as the relevant component of the
report continues to be ofsignificance. If the research report fails this
test there are two possible consequences. In the case of a problem of
considerable significance, the inadequate performance of the work
may require a public notice, so that others are not led astray by its
claims. But this is always an unpleasant business, for the good faith
of the author and the competence of the referees who passed it are
inevitably called into question. Hence it is easier, and more peaceful,
for each private user of the result simply to cut his losses, and re-
build his own work around information derived from other sources.
Formally, then, the inadequately performed work remains in the
literature, presenting a pitfall for all those who might rely on it in
the future. But the very pace of advance and change in a healthy
field of science serves to neutralize such a harmful element; if it is
not canied forward as a citation in later papers, it is soon buried in
the mass of literature, and by its title and date announces itself as
obsolete or irrelevant to later work. Thus the communal tests of
value and of adequacy by a sort of natural selection; survival is
achieved by the production ofprogeny, and a research report doomed
to sterility by its weaknesses is soon forgotten.
FflGts: I"troduetitm
We can now consider those research reports which have passed
these communal tests, in respect of their claims to be bearers of
knowledge or, more modesdy, of facts. Thus, after this lengthy
 Facts and their Evolution 18S
preparation, we arrive at the point which many philosophies of
science take for granted as the elementary, not to say trivial, com-
ponent of elaborated scientific knowledge: facts. It is not easy to
define 'fact', and this may be because the concept is so fundamental
for the sort of knowledge of which natural science is an outstanding
example. For the cumulative, progressive, impersonal, and per-
manent character of achieved scientific knowledge is intimately
related to the special properties of the products of successful
scientific inquiry. As we have seen from all our preceding analysis,
science does not begin with facts; indeed, when facts are achieved
(and recognized as such by the community) the process of achieve-
ment of knowledge in that particular area of research has passed its
most challenging phase and is settling down to a routine. Hence if
we are to explain the special character of achieved scientific know-
ledge, we must consider the concept of 'fact' more closely, and see
how it can best be understood.
There are several sorts of human experience of natural science, in
whose terms facts appear most naturally as the 'hard, massy, and
impenetrable' atomic units of scientific knowledge. Facts are the
best candidates as the bearers of certainty and truth in science; they
are exhibited to apprentice-scientists as the foundation and proof
of scientific theories; in the short term development of science they
are the court of final appeal in the judgement of theories; and in a
long term retrospective survey of a matured field of science there
will seem to be a few salient facts which stand out as survivors from
a wreckage of the erroneous theories of the past. Although the most
magnificent achievements of natural science are in its great theoreti-
cal syntheses, these must be tied down to brute facts in several ways.
Their foundations in experience must rest on attested facts, they
must predict new experiences which prove to be real and factual,
and they must survive collision with contrary facts. Finally, no
matter how hard we try to illustrate the human, creative endeavour
of scientific inquiry, or to show how bold theorizing and even specu-
lation are deeply woven into any great advance in science, we must
acknowledge that what counts in the long run in science is the in-
crease in factual knowledge of the world around us.
Yet, in spite of all this, the atomicity of facts cannot be absolute.
There has been much philosophical debate on whether any signifi-
cant assertions can be entirely free of theoretical assumptions and
entanglements; and I have already argued that the objects of the
186 The Achievement ofScientific Kn01lJledge
assertions in science are classes of intellectually constructed things
and events. This does not quite destroy the epistemological primacy
of facts in science, for one can merely require that the theory
associated with a particular fact be independent of the relevant
aspects of the theory being tested through an encounter with that
fact. Epistemology aside, practical experience and history show that
even the hardest facts are not quite impenetrable. For my analysis of
the achievement of scientific knowledge up to this point has shown
how much skill and judgement is involved in the work of creation
and assessment at every stage. It would require a repeated miracle
for those particular research reports which contain genuine facts
suddenly to identify themselves clearly and indubitably before the
scientific community. Hence the judgement of what is a real fact
must be a social one, subject to the same sorts of influences, and
subject to the same sorts of errors, as the other judgements I have
discussed. And the errors are there to be found in the archive of
scientific publication of the past. What is accepted as a 'fact' today
is at risk of being dismissed in the near future as an erroneous con-
clusion based on crude data, which was interpreted by an incorrect
theory.:& Thus the objectivity and certainty of science seem to dis-
solve before our eyes, and we are brought back to the paradoxes
with which this chapter began. For their resolution, an abstract
epistemological analysis will be insufficient; we must continue with
our description ofthe social processes ofthe achievement ofscientific
knowledge.
:a Although particular results in research reports are frequently conttoverted, the
destruction of a 'fact', IS 1 shall soon define the term, is relatively rare in matured
sciences. The more common fate is for an old fact to fade away; and those facts which are
sufficiently basic to merit refutation will usually be statements functioning as principles,
nther than directly generalizing particular experimental results. The classic example ofa
refuted fact is the principle that the orbital motions of the planets are compounded of
uniform circular motions, the so-alled 'Pto1emiac' asttonomy. This fitted common
sense, IS well IS the accepted physics of the heavens, and was the natural physical
interpretation of the mathematical tools used in any calculation (ancient or modem) of a
cyclic process. In more recent times, the 'conservation ofmass' was a basic fact ofphysical
science, from its enunciation by Lavoisier until its refutation by Einstein. A more
particular fact, only recendy refuted, is the defining property of the 'inert gases'. Their
significance is shown by this desaiption, of 1950: 'The inert gases occupy a peculiar
position in Chemistry. They are practically devoid of chemical properties, and yet for
that very reason they have provided the key to the whole problem of valency and the
interpretation ofthe Periodic Classification.' See N. V. Sidgwick, Tile CMmic.l Elemmts
tuUl t1lei, ComptnllUls (The Clarendon Press, Oxford, 1950), p. I. In 1962 there appeared
the fint reports of compounds of these gases. (I am indebted to Mrs. Rachel Country-
man for this enmple.)
 FlIlts lind t!leir Evolution 187
Facts: a Comparative Anafys;s
We may gain insight into the special character of'facts' byexamin-
ing the principles underlying their distinction from other sorts of
assertion about the external world. Consider, for example, the set
of terms: 'fact', 'law', 'theory', 'hypothesis', and 'model'. Working
our way back from the end of the list, we see that we start with very
tentative assertions; an hypothesis is something which we may fully
expect to be falsified, and a model is claimed only to have some
analogical relation to reality. At the beginning of the list are those
assertions which we make with great confidence, as 'law' and 'fact'.
Proceeding along the list in the other direction, we are starting with
assertions tied more closely to particular experience, and ending with
those which are remote from experience and are only guesses or
constructs to aid in its explanation. The most privileged position
is that held by 'law', which enjoys both a firm ground in experience
and a power of unifying and explaining. Hence it is no surprise that
for many generations the goal of natural science was to discover the
'laws of nature', even after the theological overtones of 'law', in-
herited from the earlier natural philosophy, had become muted. Nor
is it surprising that the two grounds of differentiation give the
same ordering of labels; for in general, the more closely an assertion
is tied to experience, the less likely it is to be exposed as false. How-
ever, the two schemes are not identical, and a discussion of their
differences will bring me directly to my own thesis.
In the attempts to exhibit the logical structure of completed
scientific knowledge there has been much analysis of the different
status of facts, laws, theories, hypotheses, and models in their re-
lations to particular experience. Such an analysis is not merely an
epistemological exercise, for unless the student or the practitioner of
science is aware of such differences among the assertions in his field
he will blunder into pitfalls as soon as he proceeds to the use of
existing results; I have already discussed this in connection with
'evidence'. But it has proved impossible to produce a set of formal
criteria by which all the different types of assertions are neatly
sorted out, and each assigned its appropriate epistemological status.
The reason for this is that these different forms of assertion are
taken as the species of completed, static scientific knowledge; and as
we shall soon see, such a simple schematism is inadequate for a
description of this complex and ever-changing product of the social
activity of science.
 188 The Achievement ofScimtijie Kno1lJletlge
Continuing for the moment with this traditional scheme of
classification of assertions, let us return to the first ground of
differentiation, the confidence in its survival. For our consideration
of the problems of permanence and change in scientific knowledge,
this is the more fundamental one. And we see that relation between
expected survival and closeness to particular experience is not very
tight at all. Some of the classic 'laws' of nature have been derived
from particular experience by pure or nearly pure induction; such
are Kepler's laws ofplanetary motion, and Boyle's law ofthe pressure
and density of gases. 3 Others have been produced as necessary
axioms for an elaborated theoretical science, quite remote from
particular experience either in their discovery or in their testing; and
yet by their solidity they have earned the title of 'law' rather than
'hypothesis' or even 'theory'. Examples of this class are the laws of
dynamics and of thermodynamics. And it may easily happen that
a particular set of accepted 'laws' turns out to be more solid and
enduring than those 'facts' which were adduced as evidence in its
favour when it was first announced; a famous case in point is that of
Mendel's laws of particulate inheritance.
Hence for our present purposes we do best to neglect the tradi-
tional divisions ofthe species ofthe units ofscientific knowledge 1lnd
concentrate our attention on those features of an assertion about the
external world which mark its capacity for survival. In this way we
will approach the definition of scientific knowledge; for the essence
of this is permanence, rather than a particular logical structure.
Fans: the Dejining Property
In discussing the tests on published research reports, we have
already described two properties which are necessary for an assertion
to become a fact. These may be called 'significance', in that it must
be noticed by someone ifit is not to fall into oblivion; and 'stability',
in that it must be capable of reproduction and use by others, if it is
not to be rejected as spurious. These two are necessary for an asser-
tion to become a fact, but in themselves they are not quite sufficient.
In addition, it must have a property of'invariance'. We have seen how
science advances through the investigation of problems; how each
I Kepler's painful path to discovery is an oft-told tale; for the complexities hidden
behind Boyle's Law, see C. Webster, 'The Discovery ofBoy1e's Law, and the Concept
ofthe EIuticity ofAir in the Seventeenth CcnturT, At-llliwfor Hillor1 Dfulltt Sti""",
a (l96s), 4fI-SGa.
 FflGts and their Evolution ISg
problem is concerned with the drawing of conclusions about certain
classes of intellectually constructed things and events; and how
these classes, existing only through the determination of their
properties, change even in the course of the investigation of a single
problem. When a solved problem has been presented to the com-
munity, and new work is done on its basis, then the objects of in-
vestigation will necessarily change, sometimes only slighdy, but
sometimes drastically. In a retrospect on the original problem, even
after a brief period of development, its argument will be seen as
concerning objects which no longer exist. There is then the question
of whether it can be translated or recast so as to relate to the newer
objects descended from the original ones, and still be an adequate
foundation for a conclusion. If not, then the original conclusion is
rejected as dealing with non-objeets, or as ascribing false properties
to real objects. But if such a translation or recasting is possible, then
the original solved problem is seen to have contained some element
which is invariant with respect to the changes in the objects of
investigation.
It would be tempting to try to make this notion ofinvariance more
precise, by constructing a formal scheme in which certain'elements
remained invariant in the accepted mathematical sense. But be-
cause of the subtlety of scientific arguments as well as the variety of
elements in a problem which may remain invariant through later
developments, it is doubtful that such an effort would be fruitful.
At one extreme we have those assertions which are capable (in
retrospect at least) of being derived directly from experimental data
which is cast in terms of very simple and fundamental objects of
inquiry, and so which can weather all the storms of reinterpretation.
Such primitive specimens include Boyle's law, measured by a
static force, and Black's law of the partition of heat, measured by a
temperature. At the other extreme are assertions which are not
capable of a simple test, and whose form underwent multiple
transformations and differentiation while they were still being
established; the second law of thermodynamics is an example of this
class.
It is possible for the transformations under which a fact remains
invariant to be very sudden and deep, to the point where a scientist
can recognize a 'fact' in a piece ofwork whose objects and arguments
he otherwise completely rejects. This occurs most commonly in
mathematical sciences, when a mathematical formalism is produced
190 The Achievement ofScientific Kno1lJledge
which is universally accepted as true just because it is so beautiful,
but whose interpretation is the subject of violent disagreement. One
example of this is Fourier's equation for the propagation of heat in
solid bodies, whose correctness was accepted, but for which Poisson
worked for some thirty years to provide an alternative derivation. 4
Another example is Maxwell's equations of the electromagnetic
field, whose derivation was unacceptable to any of his Continental
colleagues, and whose conceptual objects were transformed out of
recognition within twenty years. 5
Thus 'invariance', along with significance for further work and
stability under repetition and application is a necessary condition for
a component of a solved problem to be accepted as a fact; and all
three together are sufficient. I shall use this term to apply to any
statement of the properties of the objects of inquiry (thus excluding
tools) which meets these three conditions; for these are the state-
ments intended as assertions about the external world. Two sorts of
variety are thus coalesced into a single category. The first is the
different components of a solved problem which may acquire the
status of fact; not merely conclusions, but information, and perhaps
even axiomatic statements not derived from data, but serving as
part of the foundations for a theoretical argument. Also, under 'fact'
I will include statements which are qualified in a variety of ways,
falling under such categories as 'law', 'theory', 'principle', or
'model'. These distinctions are extremely important in some con-
texts. But in the present one, of the further development of the
materials of a solved problem, they are not relevant. For the accept-
ance ofan assertion as factual in this sense does not depend systema-
tically on how it is qualified when it first appears in its research
report, or on the pattern of argument within which it is established.
Again, the varied functions which facts can perform, as descriptions,
.. Fourier's work was submitted to the IDstitut de Fnnce in 1807, and received a
brief derogatory notice by Poisson in the BfI1letj. M 1IJ StKihl PlJiltmuuifl" (see Fourier,
OtllfJrtS, ed. Darboux (Paris, 1890), i, 215-21). The life-long rivalry between the two
men led Poisson to attempt to re-derive Fourier's equations from sound Laplacian
principles. In 1835 he published his Tlltont Mllt1Jnutifl" de 111 Cllllleur, where this was
done; by this time both Fourier and Laplacian physics were dead, and Poisson's
laboured treatment of the derivation had no issue. See G. Bache1ard, EtwJe "" I'ITJollUitm
tl'1IM JWobQme de PIl, . (Vrin, Paris, 1928).
I The classic statement is by Hertz: 'To the question "What is Maxwell's theory?" I
know of DO shorter or more definite mswer than the following: MaxweIl's theory is
Maxwell's system of equations.' See H. Hertz, Elenrie W.ws, tr. D. E. Jones (London,
1893), p. 21.
 Facts lind their Evolution 191

explanations, challenges, or heuristic guides to further work, are

also independent of the traditional categories.
We see that facts do not come at the beginning of a project of
scientific inquiry, but are the outcome of a lengthy process of work,
with the individual phase of the investigation of a problem followed
by the social phase of the testing, through use, of its solution. 6 Just
as the achievement of an adequately solved problem is the closest
that an individual can come to discovering the truth about the
natural world, so the achievement of facts as I have defined them
is the closest that a scientific field, at any point in its development,
can come to the same goal. The difficulty of achieving genuine facts
will vary enormously over different fields of study. In fields which
are well matured in their methods of work and controlling judge-
ments, and which are descriptive rather than theoretical in character,
facts are easy to come by; indeed, the transition from refined data to
facts may be hardly noticeable. At the other extreme are fields which
are both immature and strongly theoretical, so that the objects of
investigation are replaced with dizzying speed. In such cases, one
can find libraries full of research reports, all of it the product of
hard work and some of it of inspiration as well; but no facts.
The Evolution ofFacts: Descendant Problems
To see how some facts do survive to yield scientific knowledge,
we must first study the patterns of evolution of the various com-
ponents of a successfully solved scientific problem. When we look at
a completed research report, we find first of all a conclusion, and
with it the information which serves as evidence in its argument. To
the extent that the tools and techniques used in the work are other
than completely standard, they too will be described; for the
strength of the evidence depends on the means of production of the
data and information from which it is derived. The problem, too,
6 In framing my definition of 'facts' in terms of the survival of tests, I am of course
developing insights first argued by Karl Popper. At this point of my 'ntional remn-
sttuction' of the process of achieving scientific knowledge, the tests do not involve any
metaphysical or even ethical commitment on the part of those imposing them, for they
are applied to the works of others. But I have already mentioned, in c:onnection with
pitfalls, that a self-aitical attitude is necessary, even on narrowly prudential grounds,
if worthwhile results are to be achieved. We will later see, in connection with quality
control and ethics, that even such prudence will not be sustained in the absence of a
genuinely ethical c:omminnent shared by the members ofa scientific community. Hence,
as we shall see, Popper's deeper message, as distinct from his particular solution of an
epistemological problem, is woven into my own argument.
192 The Achievement ofScientijie Knowledge
will receive some discussion, to a degree which depends strongly on
the literary conventions within the field. The objects of the inquiry
will be mentioned, and their names used throughout the report; but
they will be discussed explicitly only if they have suffered some
significant change in the course of the work. Along with these
explicit components of the research report there are those which
remain implicit in all but the most deep and revolutionary works.
Among these are the patterns of argument, and the criteria of
adequacy and value used in the solution and the choice of the
problem. These different components of a research report (which
are of course only an abstraction from a complex and subtle reality)
go their separate ways once the report enters the public domain.
They provide different sorts of stimuli to later workers, serve
different functions, and vary in their permanence. It is through the
interweaving ofthe lines of descent of these various components of a
succession of works that scientific knowledge develops its tough
fabric, so resistant to trivial refutations and simple revolutions.
The problem itself is the most ephemeral part of the whole work.
Setting and solving a problem is a creative aet, whose accomplish-
ment depends in many subtle ways on the person doing the work, his
background and his environment; all contributing to what I have
called his style. This is the aspect of scientific work which is most
similar to that ofaesthetic creation. Although we may know far more
about mechanics than, say, Sir Isaac Newton, we cannot (even if we
wished to) reproduce his particular achievement any more than we
can that of Rembrandt. Of course, there are now some scientific
fields which are so highly developed, particularly on the experi-
mental side, that a problem-situation can crystallize in only one way,
and then we must speak not only of simultaneous discoveries, but
of simultaneous working on the same problem. But these are as yet
the exception, and the uniqueness of deep scientific creation will be
the rule so long as science admits ofgenuine growth. Even in routine
scientific work, a problem known to be solved is a dead problem; the
task cannot be repeated even if one might want to, for the world is a
bit different from what it was when the work began. Yet the problem
itself may continue to live through its descendants, in a variety of
ways which we shall now discuss.
Ifwe have an atomistic view ofthe growth ofscience, then we have
no means of understanding the lines of descent between problems.
Once a discovery has been made, or an hypothesis confirmed or
 FlIlts and t!leir Evolution 193
refuted, that unit of work is completed. One may lay one's brick
of knowledge upon the top of the pyramid, and then look for some-
thing else to do. But in our discussion of adequacy we saw that the
practice of science is otherwise. For a solved scientific problem is
not, and cannot be, a closed and perfect structure. Rather, it is the
imperfect outcome of a particular project, where the conclusion
which was drawn depended on the data which could be produced, the
tools which could be deployed, the special skills and tastes of
the craftsman who did the work, and even the first formulation of
the problem with which the intensive work began. As the final con-
clusion takes shape in the later, convergent stages of the work, the
scientist is concerned to have an argument which is adequate for the
needs ofthat conclusion; anything more is a wasteful luxury. Hence,
when the conclusion, with its argument and supporting evidence, is
presented to the community, it is necessarily rough-hewn. Those
who find the report significant will discover a variety of ways on
which the work can be carried on, by modifications of the tools or of
the argument, or by reorganizing or redirecting the work. From
these possibilities, a new set of problems can be set and investi-
gated. If they are taken up, then the original problem can be said to
have direct descendants.
These descendant problems soon mix with problems derived from
other sources; and so we speak more properly ofa 'lattice' of descent
rather than a 'sequence'. The later work produces changes in the
objects of investigation and the tools; and so the original problem,
itself an element of a descendant-lattice, usually sinks into obscurity
and oblivion. There are two main classes of exceptional cases. The
first is the 'classic' problem, where new objects and tools are created,
with such depth and power that a long sequence of descendant
problems can coast on the insights contained there. 7 Even when the
tools are transformed out of all recognition, the permanence of the
objects and of the statement of the overall problem, usually in-
soluble as it stands, justify the retention of the name of the ancestor
problem and its author for the description of the field. It is in this
sense that Newton was the father of classical dynamical astronomy,
'I In some of the mathematical sciences, it is possible to have a 'perfect' problem. C.
Truesdell desaibes it as follows: 'Every DOW and then something is done in a way that
is final; problem and solution become one, so that soon nobody Ioob at the matter in any
other way, and after a little it is taught to beginners IS a matter ofcourse.' See 'RatioDal
Mechanics of Deformation and Flow', PrDteetli",s Df tile 4t1l l"ttr1Jati01Ull Congress 011
RJuoloU Gohn Wiley & Sons, 1965).
194 The Achievement ofScientific KnollJledge
and Einstein oftwentieth-century physics. The other exceptional case
lies, in a sense, at the other extreme. In fields where data is found
rather than manufactured, so that the development of powerful
abstract theories is impossible, a grand problem can seem to remain
unsolved, and even undeveloped, after generations of work and
debate. This is most noticeable in fields which are 'history', either of
nature or of society. Of course, sometimes the appearance is true;
and a great synthetic insight, riding on the prestige ofa distinguished
scholar, can entrap generations of successors who try to isolate the
hard core of knowledge in it, and eventually realize that there is
nothing there. s But this need not be so; and the eventual rejection
of a grand problem can also come about through the growth of
knowledge which was fostered by the study of it and of its descen-
dants.

The Evolution ofTools

At the opposite extteme from the problem and its descendants,
are the tools developed in the course of its investigation. In them-
selves, they do not consist of assertions about the external world,
unlike problems or facts; and a good tool has a life (in itself or
through its descendants) independent of, and longer than, the
problem which gave rise to it. But similarity in function between tools
and facts will enable us to draw parallels between their patterns of
evolution. To be successful, a tool must first satisfy conditions
analogous to those defining a fact: it must be significant and stable,
and also applicable to problems and materials other than those
• Two such 'gnnd problems' in the history of early modem Europe are commonly
associated with the name ofMu Weber; they concern the relation of'Protcstantism' or
'Puritanism' with 'capita1ism' and with 'science'. On the former, a possible way out of
the monss bas been shown by H. Trevor-Roper in the title essay ofReli,;on, tile RefONllll-
litm tmtl S«w Cllate (MacMillan, London, 196'7), pp. 1-45. The latter thesis has
always bad to contend with the contrary evidence of the Catholicism of such important
figures IS Copernicus, GaIiIeo, Descartes, and Gassendi; and the focus for study has
shifted to England of the middle of the seventeenth century. Even there, the 'Puritan'
inftuence DOW seems to be dissolving into 'scctariaDs' on the one hand, and upper-class,
moderate 'LatitudiDuiaDs' on the other. For these, see P. M. Rattansi, 'Paracelsus and
the Puritan Revolution', AfII1Iiz, I I (1963),24-32, and B.]. Shapiro, 'Latitudinarianism
and Science', P"" flU Pre."" 40 (1g68). 16-41. A clue to the vacuity of such 'grand
problems' in history is their association of two very general entities, each of them, to the
extent that they exist at all, undergoing a complex development in time and place. One
might equally well argue for a fundamental antipathy between the 'spirit' of 'Judaism'
and those of 'capita1ism' and 'science', since during the aucial three centuries from I SOO
to 1800, the Jewish contribution to either is small.
 Facts and their Evolution
195
associated with its first creation. Once established as a genuine tool,
it undergoes evolution along several independent paths.
The most straightforward path of evolution is the refinement of
the tools used in a particular problem, for the more easy and effective
production ofthe same sort ofdata and information. The descendant
lattices of such refined tools will be closely related to that of descen-
dant problems. Indeed, the technical problems of refining tools have
a direct influence on the patterns of evolution of their associated
problems. For the choice of problems is governed by judgements of
feasibility and cost, along with value; and improvements in tools,
real or potential, will affect those judgements. There is also a more
subtle influence; for the available tools can even alter the objects of
investigation, shaping them towards a conformity with the data and
information which the tools can produce. 9 But the tools are not an
unconditioned, constant enemal force in this process; they too
suffer Clwlges in the process of their evolution. For their refinement
makes them more powerful, but also more sophisticated and
frequently more specialized. Hence they require a deeper tech-
nician's knowledge for their mastery and use, and may call into
being a separate class of tool-experts, with their ambiguous relation-
ship with their clients. Also, as the tools become more specialized,
the understanding of them becomes restricted to a decreasing group
of scientists or tool-experts, and they are in danger of becoming
esoteric and hence sterile.
This danger is generally averted through the realization ofanother
basic path of evolution of tools: their being applied to fields com-
pletely outside the desendant-lattice of their original problem. 10 For
a good tool is capable of extension to a wider class of problems and
objects, unanticipated at the time ofits first devising. We can see how
this is possible by recalling our earlier discussion of the special
properties of tools. They are usually applied to producing materials
for the prior phases of the work on a problem, as data and informa-
tion. Such materials are less specialized to the particular problem at
hand than is the evidence derived from them; and the tools, even
9 The influence of the physical equipment in determining the possibilities and direc-
tion of advance in a science is clearest in the case of astronomy; and we may note that
theory and even grand speculation have not been inhibited thereby. See A. Panneboek
A HilltW1 ofAstr0fl01tl1 (Allen & Unwin, London, 1961), especially Part 3.
10 At this point we may use the language of sociological theory, and say that a tool
survives through the eventual discovery of 'latent functions' for which it is suited. I am
indebted to Dr. J. Wootton for this point.
196 The Achievement ofScientific KnollJledge
if designed for their function in that particular problem, will be of
correspondingly wider applicability. Also, as we saw in the case of
statistical tools, information which yields crucial evidence for a
particular argument may well relate to some very general and abstract
properties of the objects of investigation; and the tools which pro-
duce such information can be extended very widely indeed. Finally,
those very special mathematical tools which provide the language in
which an argument is cast can, by their complete abstractness,
perform such a function in quite unrelated fields.
The process of extension of the use of tools comes about in many
ways. The most dramatic is when a fully developed, sophisticated
tool is seen to be capable of use at the centre of the investigation of
problems in another field: either in the production of that data and
information from which the essential evidence is derived, or as the
language of the argument itself. In such cases we can speak of the
'invasion' of a field by a new tool, its subsequent transformation
(including its objects of inquiry), and the development of new
descendant-lattices of problems and their associated tools. I I The
other extreme case is one of 'evolution' rather than 'revolution':
where the function ofthe borrowed tools is less critical in the solution
of the problem, and where neither a sophisticated version ofthe tool
nor a specialized technician is called for by the needs ofthe situation.
Only when the tool will be involved in the production of a crucial or
delicate piece of evidence is it necessary for its user to have a full
appreciation of its possibilities and pitfalls. If it is applied in a
routine fashion in the production of data, then a simple, rough-and-
ready understanding is sufficient. What is desired in such contexts is
a standardized version ofthe original tool, robust rather than refined
in its design.
Hence we should distinguish between the extension of the use of
I I The field ofcrystallography, which had evolved through the whole ofthe nineteenth

century using precision optical iDstruments for angle measurements, and the abstract
mathematics of group theory for theories of internal structure, was invaded in 1912 by
X-ray tedmiques invented by von Laue: and the field of X-ray crystallography soon
grew to become of the greatest significance in physics and biology. Materials for an
history ofthis field have been assembled in an exceptionally fine collection, Fift1 Yeus
ofx...,., Dijfrlltl;(JfJ, ed. P. P. Ewald (Oosthoek, Utteeht, 1962). In this, an essay by J.
D. H. Donnay, 'For Auld Lang Syne' (pp. s64-g) quotes a crystal10grapher of the old
school, A. F. Rogers, 'Geometrical crystallography has bad a glorious past, and it will
have a glorious future. t (I am indebted to Mrs. Rachel Countryman for this reference.)
Such invasioDs by tools are not always beneficial; particularly when an immature field
is invaded by quite inappropriate tools, its work can be seriously disturbed. We sball
discuss this in more detail below.
 Facts and their Evolution 197
the tool itself and the development of standardized versions of it. I Z
The two processes are not exclusive; for whenever a particular
physical tool is manufactured in quantity, or an intellectual tool
described in a book, it is necessarily standardized to perform a
variety of functions, operating on different objects in various fields.
We can see the extreme case of standardization in the development
of versions of tools suitable for use in teaching. Having learned
something of the craft of manipulating with this class of tools, from
experience with his 'fool-proof' models, the student will in later
years be better equipped to cope with more sophisticated versions in
the course of his scientific or technical work.
One feature of the standardization of tools is relevant to the
general problem of this chapter. This is, that a thoroughly standard-
ized version of a tool will be longer-lived than a sophisticated,
specialized version. This longevity arises from several causes. A tool
which is designed for the most effective production of very speci-
alized data or information will be more sensitive to changes in its
function; while one which performs reasonably well in the pro-
duction of non-critical material in one problem will be likely to do
so for its descendants. Also, the standardized tool, being, designed
for functioning in a wide variety of fields, will be less affected in its
overall use by changes in anyone of them. Finally, unless fashion
and prestige-replacement are dominant in a particular scientific
community, such tried and true tools, whose possibilities and pitfalls
are well known to the amateur craftsman users, will be replaced only
reluctantly.13 The extreme case of such permanence will be seen in
12 The microscope provides a good example of a tool whose use has been extended
over all fields of science and technology, and which has developed new specialized
versions and their standardized descendants. Indeed, so pervasive is such a tool, that the
writing of its history would be an enormous task, involving the close attention to experi-
mental technique in a great variety ofcontexts. The standard recent history, s. Bndbury,
The Evolution of the Microscope (Oxford University Press, 196'7), necessarily confines
itself to advances in design in a more or less linear sequence.
A picturesque but also penetrating desaiption of the evolution of tools was given by
]. aerk Maxwell, in a review of papers by Lord Kelvin in 1872: (The reader) 'may also
study, in the recorded history of electrometers, the principles of natural selection, the
conditions of the permanence of species, the retention of rudimentary organs in manu-
factured articles, and the tendency to reversion to older types in the absence of scientific
conttol.' See Tile Scimtiji& P.pers ofJames Cieri MU7I1eli (New York, 1890; reprint
Dover Publications, New York, no date), ii, 3°4-
13 A most interesting study on the cultural influences on the exteDsion of a mathe-
matical tool is A. N. B. Garvan, 'Slide Rule and Sector: a Study in Sc:ience, Technology
and Society', Proceedings of the Tmth l",eruhtmlll Convess of Kuror, of Sdnlee,
(Hermann, Paris, 196.f.), i, 397-400. The historical problem is why the slide rule,
198 The Achievement ofScientific KnollJledge
schoolteaching, where the combination of poverty and inertia can
occasionally cause the retention of experimental apparatus for
generations after it has ceased to have any sci~ntific significance
whatever.
We shall soon return to a discussion of standardization in a wider
context; but there are some features worthy of notice in all these
paths of evolution of tools. Much useful work in science is done in
the development of tools, and even in the exploratory work ofseeing
whether they can be adapted to perform new functions. Such work
rarely involves the great conceptual advances that are the most
dramatic part of the growth of scientific knowledge. This is not to
deny that such work can require great skill, talent, or even courage.
But the production or design of things for a preassigned function,
whether they be tools or particular items ofinformation, is a problem
more of a 'technical' character than scientific; and it generally calls
for a different style of work. It is on such tasks that uninspired
'scientific manpower' can be usefully employed.
But it must not be thought that all work on tools is for the unsung
and anonymous workers, no more than that tool-experts are always
in a servant-relation to tool-users. It is not merely that the first
devising and development of very powerful tools requires great
talent and commitment; but the further development of such tools
can give rise to coherent, self-contained scientific disciplines, whose
objects of investigation are the descendants of the original tools. In
such fields, the possible functions of the objects may eventually be
reduced to a minor or even negligible part ofthe inquiry, and purely
scientific problems become dominant. The tendency for tools to
become the ancestors of scientific fields is most noticeable in mathe-
matics; from geometry itself, to differential equations and statistics,
the process has repeated itselfmany times. An analogous process can
be seen in the development of chemistry, which emerged from the
status ofa tool-providing art for metallurgy and medicine to become
an independent scientific discipline. Such an evolution of a field
entails drastic changes in its methods, and in particular in the con-
trolling judgements of adequacy and value; and with them the

invented in the early seventeenth century, remained in obscurity for 200 years before
replacing the sector or 'mathematical compasses' IS the standard hand~culating
instrument. There are many facton involved, including the permanence ofcraft tradition,
the teaching of mathematics through geometry, and the retention of a 'classical' theory
of architectural design, all until the later nineteenth century.
 Facts and their Evolution 199
corresponding transformations in the character of the community
involved.

The Evolution ofFacts: Standardization

The patterns of evolution of problems and of tools present the
sharpest possible contrast. A problem is ephemeral, and lives on
only through its descendants, which become increasingly remote in
form and content from it. If a standardized version of it survives,
this will be only as a textbook exercise; transformed so as to be
soluble by routine manipulations, it will be a caricature of its
original. Tools, on the other hand, have not only direct descendants,
but also descendants from their extensions, and finally the more
permanent standardized versions of their original or later forms.
The patterns of evolution of facts lie between these two extremes;
and this could be expected from the position of science between
philosophy at one extreme, and the crafts at the other. In the short
run, an important function of facts is to pose problems or, more
precisely, to create problem-situations. An interesting fact presents a
challenge: if it is in conflict with other accepted results, one or the
other must be rejected or modified; and if it promises further
advance, it calls for its own improvement. In these ways, the estab-
lished fact is embedded in the descendant-lattice ofproblems derived
from its original; and in this path-of evolution it will survive only as
long as its context in the problem.
Those facts which continue in being long enough to become
knowledge must do so by a process of extension analogous to that of
tools: they must be seen to be relevant to problems in other fields
of inquiry, and remain so through all their changes. 14 As in the
case of tools, the process can take place by invasion, as when a
fact developed in one field forces its way into the awareness of
practitioners in another, as a significant challenge calling for a
response. The more common path of extension, however, is when
facts in one field are seen to be useful as information in the problems
in another. Such extensions are more demanding in the case of
14 The necessity of extension for the survival of results is shown very clearly by the
different fortunes of the leading fields of mathematics in the nineteenth century. Three
became extinct or nearly so: elliptic functions; invariants ofalgebraic forms; and synthetic
geometry. On the other hand, differential equations grew and survived through its
contact with physics; and the theory of limiting processes in analysis provided the basic
patterns of argument, and essential results, for analysis and its generalizations into
abstract mathematics of the present century.
200 The Achievement ofScientific KnolIJletlge
facts than of tools. For, as we have seen, tools can operate on aspects
ofthe materials of the problem which are not at all specialized to the
problem or even its field; while a fact, consisting of assertions about
particular objects of inquiry, must be capable of translation from
its original field to that in which it is being applied. Given this
difference, which entails a more rigorous selection of facts by the
processes of extension, the function ofsuch extended facts is similar
to that of extended tools. They may be used directly as information,
or as part of the explanation associated with a particular tool or
tool-subject; but in any case, they perform a function as a means to
the solution ofthe problem, and are not part ofthe evolving materials
of the problem itself.
The process of extension of facts will then necessarily involve
standardization, just as in the case of tools; and the reasons for the
longevity ofthe standardized materials are the same in both cases. It
is important to realise that standardization is not merely a sufficient
condition for the survival of a fact to become knowledge; it is
necessary as well. For if a fact remains tied to its original problem
and its descendants, it will die with them. It will stay alive longer
only if it performs a function in work on other problems in other
fields; and for this it must be available in a standardized form, for
use as information or in connection with a tool. Thus the materials
from which scientific knowledge is achieved are necessarily remote
in form from their specialized originals; they are robust, general-
purpose, standardized versions. For appreciating the properties of
achieved scientific knowledge, and the conditions for its achieve-
ment, it is necessary for us to study these standardized materials in
some depth.
Although the functions of standardized facts are very similar to
those of standardized tools, there are some significant differences
between the things themselves. One way of describing the difference
would be through the distinction between a certain class of technical
problems (for tools), and didactic tasks. These differences are partly
characteristic of the objects of the work; but they include social and
institutional aspects as well. A standardized and simplified version
ofa physical tool is in no way 'inferior' to its original; the differences
in design correspond to differences in function. But the change in
content which occurs when a fact is standardized can be interpreted
as a degeneration. For a standard fact is not only something which
performs a function in the solution of problems; it is also an asser-
 Facts and their Evolution 201

tion about classes ofthings and events, intended to relatetotheexter-

nal world. Now, as such an assertion passes into the social phase of
its development, after its original appearance in a research report,
its content inevitably changes. Even in the descendant-lattice of pro-
blems, its increased precision and sophistication will frequendy be
achieved by the sacrifice of the deep analysis which enabled its first
formulation, but which was either unnecessary or incapable of
retention in subsequent work. In particular, points of obscurity and
ofunresolved conceptual confusion in the objects ofinquiry will tend
to be overlooked, in the concentration of interest on those aspects
which are capable of straightforward further development. But
when the fact undergoes standardization, not merely the nuances of
its first intimation, but even some important but subtle aspects of
the assertion or its objects, are smoothed over and forgotten. This
seems, and may indeed be, a regrettable vulgarization, especially
when the end-product is examined by an expert in the corresponding
descendant field of research. But it is quite necessary, if the fact is to
be useful to those who lack the time, skill, or inclination to master
the elaborate theoretical context in which its sophisticated versions
are comprehensible.
We may illuminate the distinction between the standardization of
facts and oftools by considering two intermediate cases. One ofthese
is the body of 'explanation' which accompanies any tool or piece of
information. Although this will involve assertions of a factual
character, it is very strictly designed for a special function: to assist
in the mastery of skills for the competent use of the material. Hence
the requirements of sophistication, coherence, or even faithfulness
to its original, are of quite secondary importance. As we have seen
previously, even a totally incorrect theory can serve as a guide to
successful craft practice; and so an explanation of this type may be
adequate to its function in spite of containing factual assertions
which have been vulgarized out ofall recognition. On the other hand,
the standardized versions of intellectual tools which are taught
rather than manufactured, have a natural upper limit to the entropy
increase. For ifthey are to perform even simple functions adequately,
and not yield utterly vacuous materials, they must retain some
essential features of the design of their originals. We can
see this 'natural minimum' of content in the case of statistics;
there is a certain critical level of sophistication of the numerical
manipulations, and in the craft skill of operation, below which
202 The Achievement ofScientific KnollJledge
nothing but obvious rubbish can be produced by any particular
statistical tool. 15
From these examples, two importan' features of the standardiza-
tion of facts can be discerned. One is that the content of a stan-
dardized fact may decay, almost without limit; the degree of
sophistication and of faithfulness to its original which is neces-
sary for its adequate performance of its function will depend very
strongly on its use. Also, it can be seen that a version ofa standardized
fact which is good enough for one function can be quite inadequate
for another; and since any standardized fact performs a variety of
functions, it will naturally appear in a variety of versions. These two
features will be discussed at greater length, for they are directly
relevant to the nature of genuine scientific knowledge.
The Evolution ofthe Objects ofInquiry: Entropy-Increase
The patterns of evolution of the objects of scientific inquiry are
very similar to those of facts, since these are the things about which
factual assertions are made. That they do evolve, there can be no
doubt; what is meant by such terms as 'force', 'molecule', 'acid', or
even 'iron' is in constant flux. For these are intellectual constructs;
their relation to the external world is neither immediate nor certain.
Philosophical reflection on this feature of the objects of scientific
inquiry can lead to a denial ofthe possibility ofreal knowledge of the
external world; but this consequence follows only when the analysis
is abstract and over-simple. Thus, th~ insight that a theoretical con-
cept has meaning only in terms of the operations yielding its
measurement, was deprived of depth by the implicit assumption
that there is a unique and simple relation between each concept and
its associated operations. 16 When we appreciate that each object of
scientific inquiry carries with it a complex burden of meaning,
derived from its history of use and adaptation, the way is open for
showing how genuine knowledge of the external world is possible,
even when cast in the terms of such artificial objects.
15 I believed that I had been exposed to the ultimate in absurdity of the application of
statistical tools when a graduate student asked me for advice on calculating Standard
Deviations; he bad a sample of three readinp. My naivety was exposed by a colleague
(Dr. ]. Dobrzycb) who had bad a similar request, where the number of readings was
two. Neither ofus have any evidence of this technique being applied to a single reading;
but authentic instances would be welcome.

man, first announced in Tile U'"

16 This was the philosophy of the distinguished experimental physicist P. W. Bridg-
of MtHlem PA,sies (MacMillan, New York, 192 7).
There is no doubt that 'operationalism' has been beneficial as a aiteriOD of adequacy in
disciplines where concepts could otherwise be allowed to float freely.
 Facts and their Evolution 203

Those objects of inquiry which survive to become the materials of

scientific knowledge are contained in standardized facts; and we can
discuss the phenomenon of decay of information through stand-
ardization in connection with them. As the objects of inquiry are
brought down through successive transformations in descendant-
lattices of problems, and by extension, the collection of accepted
factual assertions about them becomes ever larger and more hetero-
geneous. It sometimes occurs that a change in the objects of inquiry
enables a more clear and penetrating explanation of an existing
standard fact; but the opposite effect may also occur. For the new
objects of inquiry, organized around new experiences and problems,
may be ill-suited for the explanation or even description of the in-
herited standard fact. For example, the partition of heat among
bodies in proportion to their mass and 'capacity' is easily under-
stood on the analogy of water reaching the same level in a set of
interconnected vessels; for this was the implicit model in the original
experiment. But to explain this simple experiment in terms of
'energy' or some other conceptual objects developed a century after
Black's work in the course of much more sophisticated problems,
would only lead to confusion. Those who teach the old facts then
have the choice between a diffi~lt and sophisticated 'correct'
explanation, and a continued use of the old, obsolete and incorrect
one. 17 It is because of this that the relics of old and discredited

17 When facts are retailed in an older, incorrect version, it is necessary that the students
be steered away from pitfalls which were later discovered in that version; and so the
facts as purveyed become a vulgarized version of something that never existed. For
example, a consideration of 'latent heat' in the theoretical context of the eighteenth-
century work could lead an unwary student to repeat an error common then. Q!wltity
of heat was measured by the temperature change resultant on (ts inttoduetion to a body
ofgiven mass and heat capacity; and this measure was naturally extended to the measure-
ment of latent heat. At that time, there was no reason to suppose that an experimental
result as 'I have, in the same manner, put a lump of ice into an equal quantity of water,
heated to the temperature 176, and the result wast that the fluid was no hotter than water
ready to freeze' could not be ttansformed into • general measure of heat of fusion, IS
'But this quantity ofheat (which disappeared in melting the ice) would be inaeasingt by
143 degrees, the heat of a quantity of water, equal in weight to the ice alone'; and then
applied to 11"1 substance, as ' ••. the heat which any given quantity of water loses upon
being frozen-were it to be communicated to an equal weight of gold, at the tempera-
ture offreezing, the gold, instead ofbeing heated 162 degrees, would be heated 140 x 20=
2800 degrees or, would be raised to a bright retl1Jellt.' The first two quotations are from
Joseph Black, Lea"es on the Elements ofClJemistr] (Edinburgh, 1803) i, 124-5; the third
is from Rumford, 'An Inquiry concerning the weight ascribed to Heat', PlUl. TrlltlS.
(1799), p. 193. I am indebted to Dr. D. S. L. Cardwell for this point; he has used it in
his teaching to show that even pure 'phenomenological' physics can have its pitfalls.
204 The Achievement ofScientific KnollJledge
theories clutter the teaching of science, not only in the schools, but
wherever it is being conveyed in a standardized form as atool-subject.
But to attempt to clean up all this confusc!d mass, and to purify the
material by deriving it directly from its basis in modem research,
can lead to a syllabus which, however satisfying intellectually to the
teacher, leaves students in a state of confusion worse confounded.
Hence those old objects of scientific discourse which are still useful
for the description of standard facts, will remain alive even after they
are obsolete for the purposes ofscientific inquiry. And because ofthe
complexity ofthe development of those objects, as well as the variety
of teaching situations, it is impossible to ensure that every student is
brought forward in a neat progression from convenient old errors,
towards powerful new truths. More commonly, the different sets
of objects in a family will coexist, peacefully among teachers and re-
searchers who each use only one set, and confusedly among
students. IS
It is sometimes possible for a field to be 'unified', so that its
II In the teaching of elementary chemistry until recendy, chemical combination was
explained by • theory of 'valency' descended directly from the 'dualistic' theory of
Berr.clius in the early nineteenth century, and given • convincing rationale in terms of
electron structure and the electrically neuttal state of. complete molecule. It was then
difticult for students to grasp the Ceo-valence' of organic compounds, to say nothing of
diatomic molecules IS Oxygen gas. Little did we know that the same difficulties were
present in chemistry in the second quarter of the nineteenth century; that they were
partlyresponsible for the decay in theacceptanceof Avogadro's hypothesis; and that our
'basic'theory was the relic of a grand synthesis that had failed. The influence ofa subse-
quent research ttadition on the formulation of. standard fact can be seen in the case of
Boyle's Law. It was originally expressed in terms of a proportionality between the
'spring' of an elastic fluid, and its density; and we may translate 'spring' into 'pressure'
without too much violence to the original. But in modem times 'density' has been
replaced by 'volume't which in the context of the original law is not the fundamental
physical factor involved. It seems likely that when • more general gas theory developed
in the nineteenth century, with volume IS. significant parameter, the version of Boyle's
Law UICd in elementary teaching was modernized, with. consequent loss of clarity. It is
even necessary, on occasion, for the doctrine of • science IS taught to ignore certain
reliable phenomeDa, which may once have been of great interest, in order to maintain
• tidy structure of theory and ezperience. It might be expected that this will occur in
optics, where the student's eye is involved in the experiments. In Oplies, tile Seien&e of
Yilitm, V. Ronchi cites the example of. simple convex leus with an object placed at the
distant focus; the observer should see virtual images behind me lens at an infinite
distance. No one does, but this does not hinder the teaching of simple lens theory. In
more advanced wave optics, one learns of the position ofimages cast by concave spherical
mirrors; experimental tests of these theories lead to the most astonishing results, Done
fitting the theories. Yet the phenomena of highly curved reflecting surfaces were the
IUbject of great interest from antiquity to the seYeDtecnth century, and it.appears that
they bad to be suppressed in the modem period in the interest of. tidy and plausible
aplaDation of refnction and reftcctioa. (See pp. ~ 133-41.)
 Facts and tneir Evolution 205

various objects of discourse can be transformed into mutual c0-

herence, or rejected. Such unification usually takes place at the
highest level, either summarizing a completed research tradition, or
creating a coherent field out of previously diverse problem-areas.
But even in such work, one cannot escape from history. That which
is being unified was previously diverse, its problems and objects
deriving from different sources in experience. In the programme of
unification, a small set of these experiences, and their associated
objects, are taken as fundamental, and the others are exhibited as
being capable of derivation from them. But it is impossible to make
a perfectly neat structure out of such disparate materials; and then
those who teach the 'unified' science must cope with, or conceal, the
awkward unconformities at the boundaries. This phenomenon can
be seen in mechanics, which of all physical sciences seems the most
naturally unified, and derivable from a few basic axioms and ex-
periences. Its sources in experience come from diversity of practical
tasks, and several crucial phenomena. For the tasks, we have such as
load-lifting, balancing, falling of bodies, projectile-throwing, and
overcoming a resistance by impact. The phenomena include the
pendulum, and the transfer of motion by collision. These latter
served as the foundation of dynamics in the seventeenth century,
through their use by Galileo and Descartes respectively. Newton
grafted a general idea of 'force' on to the collision phenomenon for
his mechanics; while Huygens and Leibniz generalized from the
pendulum to 'vis viva', the ancestor of kinetic energy. The unifica-
tion of mechanics, started at the formalistic level by Legrange, and
deepened in the nineteenth century by the new concept of 'energy',
still leaves these separate roots in existence. Thus, Ernst Mach's
programme of banishing the concept of 'force' from mechanics as
an anthropomorphic relic leaves the field of statics as 'the science of
nonexistent motions'.19 The lifting of a load at constant velocity, as
by a water-wheel, can be explained in terms of work and energy
but it coheres not very well with the Newtonian concept of force and
acceleration. 2o And certain paradoxes ofimpact which were common
19 See Tile Sciente ofMe,"',,;,s, 5th English eel., from 7th German ed. (Open Court
Publishing Co., La Salle, IlL and London, 1942), ch. I, section v, pangnph 6 (pp. 95-6).
Mach defines 'force' as 'any circumstance of which the consequence is motion'; DOtes
that these may be conjoined as to result in no motion; and says 'statics investigates what
this mode of conjunction, in general terms, is.'
20 On the inapplicability of Newtonian dynamics to the problems of analysis ofpower

machinery in the eighteenth <:cntury, see D. S. L. Cardwell, 'Early Development ofthc

206 The Ae1Jiewmml ofSeimtijie KnolIJletlge
in the sixteenth century, such as the radically different effect of a
hammer resting on a nail and the same hammer dropped on it from
a height, are omitted from modem syllabuses, doubdess because of
the confusion they would cause in immature minds. 21
In general, the standardization of the materials of a field necess-
arily kills them. What was created in a succession of turbulent and
frequendy confused waves of advance, and then abandoned before
being fully clarified when the frontier of research shifted elsewhere,
must now be given a tidy organization. It must be presented as ifit
all started with the discovery of an elementary fact by elementary
techniques, and then grew by a linear, logical development with the
injection of new elementary facts at the appropriate places. It
matters less that such a presentation destroys the actual history,
than that it implicidy purveys a false and deadening picture of
scientific knowledge and of its achievement, and presents its
materials in truncated and vulgarized form. 22
Because of all these effects in this standardization of facts and
objects of inquiry, it is possible to speak of a decay of information,
or an increase of entropy, accompanying the process. Z3 This should
not be surprising; the physical transmission of any 'signal' involves
an admixture of 'noise'; and an entropy-increase in the manual
copying of texts was a universal phenomenon, which now enables
the reconstruction of descendant-lattices of medieval manuscripts.
A similar effect has been observed in encyclopedias, ancient and
modem." Nor should we be surprised if the increase in entropy is
Concepti of Work, Power and Energy', Brilisll """'1 for tile Hill"" of Snm&t, 3
(196'7), ~ especially pp. 212-14-
:&1 This paradox can be put quite vividly, in terms of a bow and arrow, first put in

tension with the arrow's point apiDst • plank of wood (making only a slight mark on the
surface), then withdrawn still extended, and released. The arrow will penetrate the
wood; why? Attempts by pupils (and teachers) to solve the problem shows clearly how
their training in elementary mecbanics is confined to manipulation of standard, safe
examples. This thought-aperiment was cited by Leonardo cia Vmci, StleaitmS from tile
Not,lJoob ••• , eel. L A. Richter (Oxford University Press, 1952), p. 66; and Galileo
strunled ftinly to resohe the paradoxical properties of impact forces (see S. Moscovici,
'Ilemarques sur Ia dialogue de Galilie, de Ia force de Ia percussion',RI'DIIe tl'HilloU-t des
Sne.,,, 16 (1963), 97-137.
U E. A. Carlaon, Tile Gme: II Criti&1Il Hillor, (Saunders, Philadelphia, 1966), displays
this process particularly weU. In his Jut chapter, 'Historical Conclusions' (pp. 244-58),
he asserts that the 'beads on a string' version of chromosomes was 'never more than •
straw man'.
23 I am indebted to Dr. D. S. L. Cardwell for this and many other features of the
eYOlution of facts.
M On the dcpneration of the encydopedic tradition in Rome, see W. Stahl, RtnIIIIII
 Facts and tneir Evolution 207
frequendy greater than might be considered strictly necessary for
the particular function of the standardized fact; for the natural
pressures of a didactic situation will produce such a tendency. Also,
the person involved in the didactic task of standardizing material
downwards from one level to another will frequendy lack a sufficient
mastery of the material at the higher level for a good presentation at
the lower.
It is easy to single out schoolteachers as the prime targets for a
critical analysis of standardization. Their social isolation from
research activity cuts them off from the~ stream of sophisticated
discussion of the facts and their objects; and in any event their
standardized versions of facts will usually have little similarity to
those used in current research. Hence they are generally forced to
retail standardizations of standardizations, or vulgarizations of
vulgarizations, as the case may be. These inherent limitations of the
schoolteaching situation, along with its function of imparting basic
craft skills rather than 'understanding', must be recognized if there
is to be any fundamental improvement in its quality. However, the
process of standardization of information and facts, like that of tools,
goes on at all levels up to the highest; any publication other than a
pure research report involves standardization. To do the work well
requires a special skill, which is. rewarded in the lasting success
of a classic textbook and its descendants. To do the work badly is
also possible and easy at all levels; vulgarization and incomprehen-
sibility can be found even in the most advanced and specialized
monograph literature.
At some time in the future, it might be possible to enliven the
teaching of science by re-creating parts of the history of which the
standardized materials are the relics. 25 By such a means, students
could share in the excitement ofdiscovery and conflict which attended
all the great achievements, and appreciate that the apparently tidy
organization of the material is only a means of making it compre-
hensible for its function as a tool. But the difficulties in the way of
such a programme must be recognized. It is not sufficient merely to
have good historical studies (of which not many exist as yet); but the
reconstruction of the context of discovery and debate would involve
Sdmee: Origins, Development "nil l"jlumee to tile LIIter MiIltJle Ates (University of
Wisconsin Press, 1962). A modem analogue is discussed by H. Einbinder, Tile Mflh of
tile Brit""";c,, (Grove Press, New York; MacGibbon, London, 1964).
25 A model for such an approach is V. Ronchi, Optics, tile Sdmee ofYisitm, which I
have already discussed.
208 The Achievement ofScientific KnollJledge
teaching some amount of quite obsolete and erroneous technical
material, which could be laborious and also confusing to the weaker
students. It would seem that such an approach to the teaching of
science will be restricted to a very few convenient examples, until
there are deep changes in the tools and techniques of teaching. One
can imagine a separation ofthe two aspects oflearning, with students
mastering manipulative skills privately on a machine, and then
having open-ended discussions with the teacher on the under-
standing of the material at all levels, technical, historical and philo-
sophical. But for this, neither the machines, nor the historical and
philosophical materia1s, nor the t-eachers, are likely to be available
for a long time to come.
 7
THE SPECIAL CHARACTER OF
SCIENTIFIC KNOWLEDGE

THE materials which comprise scientific knowledge are facts, of a

certain sort: those which have survived the processes of testing and
transformation, so that they remain in use, and hence alive, long
after the disappearance of the problem which first gave rise to
them. We have already seen that these processes involve losses as
well as gains; in particular, the standardized facts and their ass0-
ciated objects of inquiry are liable to suffer a loss of their content
of information, and to be vulgarized. This may well be regrettable,
but it is inevitable. No one (except possibly the historian) has any
reason for preserving a significant fact in its pristine form; the facts
stay alive only as long as they are useful in new contexts, and if they
cannot be put into a convenient form for their limited functions,
they are soon discarded and forgotten. Although such standardized
facts comprise the body of scientific knowledge, it is not simply a
collection of atomic units of hard, long-lived facts. Indeed, its
special character results from the complexity and interconnected-
ness of its materials, as they evolve through the complex and
fallible social processes of their use and adaptation. I

Different;at;on ofStandardized Facts

When we examined the standardization of facts, we observed that
this is done in order that the existing fact can perform one or several
new functions, as a means to the solution of new problems. In
general, there will be a great diversity of such functions: for the
fact can be applied in a variety of different fields (each requiring
I E. T. Bell, Tile Developmmt ofMllthntulti&S, OODleS close to my conception when he
speaks of 'residues' of epochs: that which survives, having been sublimated from a
mass of detailed results, to remain alive in the work of future epochs (p. 33).
210 The A,hievtmmt ofS'imtiji& Knowledge
its own translation of its objects of discourse); it can serve as in-
formation, or as part of the explanation of a tool; and in each
possible field and type of use it is capable of being applied at
different levels of sophistication. The number of conceivable func-
tions for a standardized fact will then be very large; and although
each particular version will perform well over a range of functions,
there is still room for a multiplicity ofversions ofthe same standard-
ized fact, differing among themselves and all remote from their
original. Thus standardization does not exclude a rich differentia-
tion; and the more basic the fact, the greater the degree of differen-
tiation. Hence when we speak of a particular fact as having survived
to become a part of scientific knowledge, we should mean a family
of particular versions of a fact, mutually related by a complex lattice
of descent from their original, and all still in flux.
This variety escapes systematic attention because it can usually
be ignored in practice. The research worker in science has a limited
range of functions for any tool or fact within his competence; and
the philosophers of science have hitherto generally assumed that a
fact, once it has emerged from the murky, irrational phase of crea-
tion, takes on a permanent, unique, and rigid form. Only teachers
find the multiplicity of doctrine to be troublesome, and then only
sometimes. And in any event their experience is not considered
weighty evidence in the socially more exalted realms of science. But
when teachers must cope with the misunderstandings and mis-
information inherited by their pupils from earlier instruction, or
tty to agree on the standardization of materials designed for differ-
ent functions (as 'units' in physics), or even teach their own material
as tool-subjects for students in other fields, the difficulties, peda-
gogical and sometimes social and institutional as well, can be acute.
The debates which ensue in such situations are frequendy vitiated
by the assumption that there exists a unique, perfect, and true ver-
sion of the facts in question (and even that this perfection is to be
found among the standardized versions in dispute).z But the case
a Emst Mach detected this tendency thus: 'Until CODtemponry times, however,
workers in this field teem more or less UDmDlCiously to have sought for a natural
meuure of temperature, • real temperature, a kind of Platonic Idea of temperature, of
which temperature read from thermometers is only an inmmplete and imperfect ex-
pression.' See E. Mach, Die PrinrJp;m tier W4rmelelwe(Leipzig, 18g6); English transla-
tion of this section by M. J. Scott-McTaggat and B. Ellis, in B. Ellis, BIUi& CtnJ&epts
.fM'''''''ntItfII (Cambridge Uni~ty Press, 19(6), p. 190. I am indebted to R. G. A.
Dolby for this item.
 The Spe,ial Clulrfl&ter ofS'imtiji& Knowledge 211

is otherwise; we have only a· variety of functions, and'criteria of

adequacy of the performance ofthe functions by the materials ofthe
various designs.
For those who have not had this unsettling experience of dis-
agreement over the obvious, it may be hard to imagine that the
fundamental facts of science are capable of such obscurity and
variety. For such facts, and their objects, imparted as part of the
most elementary craft instruction, tend to become part of the un-
questioned and usually unquestionable common sense of a field.
Thus, when C. P. Snow bemoaned the deep ignorance of the 'arts'
graduate concerning science, he used as the test of elementary
scientific literacy, the question 'What do you mean by mass, or
acceleration ?'3 Completely encased in the scientist's common sense
of his generation, he was not to know that the answer to the second
question lies deep in the mysteries of the calculus; and that on the
first question there have been as many opinions as philosophers of
physical science. The first, natural reaction, on the discovery of the
variety and obscurity in this elementary, common sense knowledge,
is to search for the true version in a sufficiendy thick book; and the
second is to attempt to sort it out for oneself. For, the thicker the
book that one examines, the more advanced the material, and
(usually) the less illumination one will find on the elementary con-
cepts. Also, each thick book will put the elementary material to a
different use, and so will offer a different version of it appropriate
to its use in the discussion of a special set of advanced problems.
The bold person who strikes out on his own to discover the unique
true meaning ofa fact can sometimes achieve new results concerning
this elementary material, which then become a part of it for the
future. Mach's analysis of the foundations of Newtonian dynamics
is a famous case in point. But genuine scientific knowledge is too
complex and rich to be comprehended in any single schematism,
didactic or analytical; the task of understanding what has been
achieved is as never-ending as the task of extending those achieve-
ments to new realms. 4
3 C. P. Snow, Tile TIN Cd"" IIfIIl lile Stinllifie RI'DDIIIIUm, 'I now believe that if I
had asked an even simpler question-such IS, What do you mean by mass, or by
aa:eleration, which is the scientific cquift1ent of saying, CII" 7011 reflll?-not more than
one in tell of the highlyedue:ated would have felt that I was speaking the same Ianguap'
(pp. 14-1 5).
.. Studies of the variety in interpretatioDs of the two !DOlt buic c:onc:eptI in ph'"
have been made by Max JIIDII1er; ICe his CtmeepI, ofFortI (Hanard Uni~ty PreIs,
212 The Achievement ofScientific Knowledge
Obscurity lit the Foundlltions ofScientific Knowledge
Another property of standardized facts, closely related to their
differentiation, is the obscurity of their objects. Those who have had
to grapple with different forms of the same standardized fact have
learned that behind the facade of simple definitions and plausible
experiences there is a bewildering variety of interpretations and
meanings. 5 Sorting these out, either conceptually or historically,
leads into a labyrinth of arguments, in which the unique, clear and
distinct definitions of the objects are impossible to locate. The best
that can be achieved in practice is another, more plausible or more
coherent standardized version, better suited for the particular
pedagogical task at hand. These failed attempts do not receive
systematic attention, because hitherto there has been no philo-
sophical thesis for which they could function as evidence. The ideal
of science as demonsttative knowledge excludes the possibility of
the obscurity of the very objects of the demonsttation; and in the
dominant ttaditions of the philosophy of science, such an obscurity
1957) and Ctmeepts ofMIUS (Harvard Univenity Press, 1961). The first chapter of the
earlier book, and the Introduction of the later one, provide arguments and examples in
agreement with the thesis advanced in this section.
5 The obscurity at the foundations of Newtonian mechanics was recognized as a case
for concern by several German scientists in the later nineteenth century. The most
famous product of this amcern is Ernst Mach, Tile Stien&e of MecluJ,,;cs: a Critical
MUl HisttJrical .Accotmt of its Developmmt (1St German edition, Leipzig, 1883; 1st
English translation, Open Court, La Salle and London, 18g3; many reprints since then).
Of equal significance was H. Hertz, Tile Pritl&iples of Mec/umics Presented in a NeJl)
FfJNII (original German edition, 1894; English translations from 18gg; reprint, Dover
Publications, New York, 1956). In the first part of the Introduction (pp. 1-14) Hertz
discusses the problem of obscurity, as it is glossed over in the teaching ttaditions, in an
illuminating way.
An elementary introduction to the obscurities at the foundations of mechanics can
be obtained by consideration of the five symbols: F=". X a. 'Acceleration' is a concept
best left to the mathematidaDs; the 'times' brings in the problem oCthe multiplication
of the measures of dlfrerent physical quantities; 'mass' and 'force' are capable of
cWrerent, interrelated definitioDS; and 'equality' is capable of sevenl very distinct
meanings. An illuminating summary of these is found in N. R. HaDson, Patterns of
DisctJVtry (Cambridge Univenity Press, 1958), pp. 99-100•
The obscurity in the mnccpt of 'image', fundamental for physical optics, is exposed
and discussed by V. Ronchi, Optics, tile Stien&e of Vision. It is not merely a question
of a 'point_image' of a point-object being an idealization, since every spherical reflect-
ing or refracting surface produces a 'caustic' curve with a cusp, and that every finite
aperture produces a spread by ditrraction effects. More than this, is that the fact that
what is 'perceived', even by a photo-eensitive substance as well as the eye, will be an
interpretation, sometimes extreme, of the pattern of radiant energy incident upon it.
To speak of 'seeing the image' created by a lens system is to be immersed, unwittingly,
in the obscurities responsible for the triumphs of physical optics. See pp. 265-71.
 The Special Character ofScientific Knowledge 213

would destroy the claim of science to be knowledge of any worth-

while sort. 6
Indeed, the claim that the foundations of science, in its con-
ceptual objects, are swathed in obscurity, immediately raises troub-
ling paradoxes. If, for example, the objects of mathematics are
essentially obscure and incapable of definition, then what guarantee
do we have that the sciences using mathematical arguments give
valid conclusions? In practical terms, if the obscurities in the
differential and integral calculus are ineradicable, how do we know
that the bridges will not suddenly fall down because of hidden
fallacies in the mathematical arguments whose conclusions have
determined their design? We can remove some of the force of this
paradox by observing its similarity to those arising from the in-
conclusiveness of scientific arguments, resulting from the absence
of formally valid patterns of argument. But this aspect of scientific
knowledge has somewhat different effects on its development than
the necessity for socially improved criteria of adequacy on the
solution of problems, and is worth considering on its own.
The present thesis can be argued by extension from one well
known to philosophers: that the basic categories of our experience
are incapable of precise definition and unique analysis. Concepts
such as 'cause,' 'change', and the like, have in them an inexhaustible
supply of subdeties and ambiguities as material for philosophical
inquiry.7 Even basic concepts which are manipulated in scientific

6 The faith in the essential clarity of mathematics was most beautifully expressed by
Fourier. Speaking of the 'analytical equations' of Descartes, he said: 'There cannot be
a language more universal and more simple, more free from errors and from obscurities,
that is to say more wonhy to express the invariable relations of natural things.' On
mathematics in general, 'Its chief attribute is clearness; it bas no marks to express am-
fused notions.' Tile Aulytical Tlleory ofHeat, 1822, tr. A. Freeman, 1878; reprinted
1955 (Dover Publications) p. 7. A mntrary impression mmes from Karl Menger:
'''Variable'' undoubtedly is among the most frequent nouns in the mathematico-
scientific jargon and hence one of the most successful words ever created•.•• Aetua1Iy,
few scientists seem to have given any thought to the problem of what variables are, and
still fewer of those who use the term formulate clear and satisfactory answers when the
question arises.' See his 'Variables, Constants, F1uents', in C",mt Issues i" tile Pllilosop"y
ofScience, cds. H. Feigl and G. Maxwell (Holt, Rinehart, & Wmston, New York, 1961),
p. 30+ This latter example indicates that mathematics contains obscure objects other
than those which have received attention in the dominant ttadition in 'foundations of
mathematics'.
, The pandoxes of change, and also of aggregation, propounded by Zeno of E1ea,
still exercise philosophers. See H. D. P. Lee,Zmo ofElea (Cambridge University Press,
1936), for tests and aualysis; and for a review of some recent controversies, A.
Grilnbaum, MotlmJ Science and Zeno's P",lIIlozes (Allen & Unwin, London, 1968).
8-s.IC.S.P.
214 The Achievement ofScientific Knowledge
argument, such as 'number', defy all attempts at conclusive ex-
planation. For what we have for analysis are intellectual constructs
derived from very deep aspects of human experience of coping with
the external world. Our practical command of them has developed
through millenia of experience; and no single formalization can
capture that body of completely tacit inherited knowledge.
The ordinary objects of scientific discourse are more specialized
and artificial, and in that measure are more susceptible of intellec-
tual analysis. But as they appear in an argument, they are mixed with
these more fundamental, unanalysable concepts, and it is not
possible to give a clear specification of their meaning relative to
neady defined alternative meanings of the basic concepts of ex-
perience. 8 More important, however, the investigation of the
scientific problem in which they are used or created does not have
the goal of an explication of their meaning. Although the con-
clusion of the argument concerns properties of these intellectually
constructed objects, the properties sought for are those which relate
to experience ofthe external world; for the ultimate goal ofscientific
inquiry is to advance knowledge of that external world. A problem
solved by conclusions about the objects as intellectual constructs
is a philosophical one. The two sorts ofarguments will have different
methods, and admit different sorts of evidence. Although there will
be borderline cases, the distinction between the two sorts of prob-
lem was well expressed, for mathematics, by C. S. Peirce: Mathe-
matics is the science which draws necessary conclusions, while logic
is the science of drawing necessary conclusions. C)
• For example, 'electric current' bas 'operational' definitions through a variety of
effects (the two most important being electromagnetic and elcctrochemica1), each one
of than measured by devices capable of precise specification in terms of elaborated
theories. Although. penetration to the foundations ofany ofthese will reveal obscurities,
these foundations are well hidden beneath the superstructure of accomplished scientific
knowledge required for the framing of the definition of 'current'. It is only in the
determination of the values of the 'fundamental c:oustants' that experimental technique
must reckon with conceptual difficulties.
In his classic monograph, .A" .ACltnml oftile Prifleipks ofM'lIsuremmt lind CllkfJlltitnJ,
N. R. Campbell observes 'the-most acx:omplished physicists are apt to flounder when
plunged into • discussion arising from general principles of measurement; international
amunittees, cbarged with the definition of standards, do not seem (to me at least) to
display that easy mastery oftheir subjects which is to be expected from the conttibution
of their members to original1eaming' (p. vi). For an example of the complications from
which the international committee on electrical units did not quite exttie:ate itself,
see pp. 131-3.
9 See C. S. Peirce, 'The Essence of Mathematics', in ESSIIYs in tile Philosophy of
Snmee, ed. V. Tomas (Liberal Arts Press, New York, 1957), p. 266.
 The Special Character ofScimtifi& Knowledge 21 5

Corresponding to their function in the solved scientific problem,

the objects of inquiry will be subject to particular criteria of ade-
quacy in their creation and development. Provided that those pro-
perties which are used in the argument do not lead to incoherence
or inconsistency in it, they are adequate to their function. Like all
other components of the solved problem, they are only as refined as
they need to be. Later investigation, either scientific or philosophical,
may reveal hidden ambiguities or incoherences in the terms used in
the argument; but whether these constitute pitfalls depends on
whether the original argument is then seen to be vitiated by these
defects in its objects. Such pitfalls are not inevitable; indeed, they
are the exception. For the conceptual objects with which the
scientist manipulates have strong analogies in their use with tools.
Those properties of a tool which are used in its application to
scientific inquiry will generally be crude and over-simple, compared
to those which have been elaborated by experts in the practice or in
the theory of the tool. Yet the tools will be adequate to their func-
tion, for their use is on materials other than those peculiar to their
own constitution, and in general they will be used on the less
specialized aspects of those materials. Similarly, the objects of a
scientific argument, whose concealed obscurities contain a wealth
of material for philosophical analysis, are used as the components
ofan argument intended to relate to something other than the objects
themselves, namely the external world. Hence a more rough and
ready conttol of the properties of those objects will be adequate for
the argument.
It is for this reason that the obscurity in the objects of scientific
discourse is generally irrelevant to scientific work of any sort. Con-
tact with this obscurity is made in particular ways, depending on
the work being done, and (in matured fields) a practical resolution
can be achieved in every case. Thus, the research scientist will
usually be aware of obscurities, at the technical level, in new
materials in a rapidly developing field. But as the field stabilizes,
there is achieved a sufficiently clear and univocal understanding of
these objects for them to be manipulated in arguments with com-
mon agreement on their meaning and an absence of pitfalls. The
concealed obscurities in the standard information and tools taken
over for use in problems are beyond the practical concern, and
hence beyond the awareness, of the research scientist. The teacher
must occasionally grapple with obscurities in this standard material;
216 The Achievement ofScientific Knowledge
but his task'is to render it plausible and easy to manipulate by routine
techniques; and so he will naturally tend to minimize or gloss over
unsolved problems at every level from the technical to the con-
ceptual. It is only when a drastic novelty of some sort is intruded
into the work, and existing routine craft skills disrupted, that
scientists will become painfully aware of the obscurities at the
foundations of their knowledge. to
Only a person with a keen philosophical awareness will attempt
to explore and resolve the obscurity in the conceptual objects he
handles. This can be of great benefit to teaching, and to research
as well, as long as this sort of philosophical problem is tteated as a
subsidiary problem in the general task, and governed by criteria of
adequacy which are not too severe. Otherwise it can have a harmful
effect on the main work; for the different sorts of problem, with
their very different functions, have quite different criteria of ade-
quacy appropriate to their solutions. Thus, an exposition of material
in physics or mathematics which derives it rigorously from very
deep foundations can provide intellectual satisfaction to the teacher,
but leaves his pupils in a state of complete bewilderment. Also, as
a field of philosophical inquiry, the examination of particular
scientific concepts present peculiar difficulties, analogous to those
of methodology. For while the arguments must meet the standards
of rigour for genuine philosophical debate, there is an embarrass-
ment of detailed data, some of which must be used for evidence,
arising from the practice of the scientists who actually use the con-
cepts. This sort of inquiry has social difficulties as well, for its
position on the borderland between science and philosophy calls
10 A striking example of this at the present time is the problem of oonverting measure-
ments in Imperial units to the S7ste- Inttrrllllitnull. For this to be done properly,
there must be a rea»gnition that a stting of digits is only an estimate of a continuous
magnitude (or olan aggregate ofindividuals which have not been munted individually);
and that the conventions Cor 'significant digits' are a tOM whose meaning depends on
the function of the digits. It is not surprising that mathematical social scientists are
unaware of this subtlety; but it is interesting that engineers, who generally have a good
aaft knowledge of tolerances, find themselves in considerable mnfusion when this
aaft knowledge is required to be adapted to the new problems of metrication. On this
situation, see K. J. A. Brookes, 'Can You Metricate?', E",;neeri", (6 February 1970),
pp. 13~. (I am indebted to Mr. B. Hunter, of the Department of Civil Engineering,
Leeds University, for this reference.) Even the author of this very perceptive paper does
not recognize the deepest obscurity of all, that of the symbol 0. For it functions both
as a filler for an insignificant place (in a number greater than unity), and also as a symbol
of a significant zero-value in a place. The ambiguity is revealed by an expression as
°
420,000; which is the last significant digit? This ambiguity in the may be one reason
for its very late invention.
 The Special Character ofScientific Knowledge 217

for a double competence in its practitioners; and our educational

system provides only very few who are so qualified. For all these
reasons, inquiry into the obscurities in the objects of scientific
discourse is a very hazardous affair. The philosophically-minded
scientist or teacher might believe, at first, that a deeper understand-
ing of a basic concept should be easy to achieve by extension of his
scientific knowledge and perusal of the relevant literature. But,
paradoxically, it is easier to become skilled in the techniques of
discussing a question like 'What is mind?' than one such as 'What
is mass?'

With this appreciation of the obscurity at the foundations of

scientific knowledge, we can see how programmes for teaching
absttact sciences by starting with their 'foundations' suffer from a
double fallacy. The first concerns the learning process itself; it is
tacidy assumed that the mastery of craft skills of manipulation is
irrelevant to true understanding. And the second is that there exists
a set of clear and simple foundations, whose understanding is
sufficient for the mastery of all their consequences. Both these
fallacies derive from a naive faith in the nature of scientific know-
ledge, which can be called Cartesian since its most influential
formulation derives from the early, optimistic period in Descartes's
career. This is, that ttuth is a simple thing, derivable by logic from
clear and distinct basic ideas. For Descartes, learning the truth is a
simple matter of proceeding by instantaneous steps through its
structure. The sort of knowledge which requires practice is that of
a skill, as playing an instrument, and has nothing to do with the
possession ofttuth. 11 Descartes's own programme derived from his
desperate commitment to locate a basis of certainty for knowledge
and morals. The modem descendants derive partly from a laudable
desire for intellectual honesty in teaching; and also from the wish
to introduce the materials and methods of sophisticated research
earlier into the curriculum, for the wider diffusion of its approach
and for the indoctrination of potential recruits. In fact, any
J J This distinction between skills and knowledge was fundamental for Descartes's
philosophy; it is announced at the very beginning of his first philosophical work, the
Re"dtle IIIl Direaionem ["genii; see Tile Philosophic.1 Woris ofDesc,'ts, ed. Haldane
and Ross, vol. i (Cambridge University Press, 1911), pp. 1-2. Descartes's conclusion
was that to discover bUth, one should not first study particular sciences, but ntber
'increase the natural light of reason'. When one is so prepared, the appropriate style
for the study of particular problems is to mnsider them in as abstract and genenla
fashion as is J*S1Dle.
218 The Achievement ofScientific Knowledge
particular set of'foundations' chosen for teaching purposes will be an
arbittary selection, and will necessarily be purveyed in extremely
vulgarized form. And students who are raised on a diet of absttact
arguments and necessarily trivial problems on their materials, will
emerge without either the skills or the general knowledge to be
anything other than research technicians in that specialized field.
If the problems of dispelling the obscurities in the objects of
scientific knowledge are truly insoluble, in principle as well as in
practice, we must consider two questions: first, whether there is
any point whatever in making an attempt at such a task; and second,
why indeed don't the bridges fall down. The second problem is
easier to solve, for the materials are already to hand from our dis-
cussion of tools and the analogous properties of the objects of
scientific 'discourse. Of course, bridges do sometimes fall down, but
in nearly every case the failure can be explained by some more
sttaightforward, or vulgar, cause. The design can be shown to have
been faulty, in failing to reckon with some feature of the technical
problem: or there might even have been too much sand in the con-
crete. The study of failure of a structure never leads to the question-
ing of the basic laws of mechanics that were presupposed and used
in its design, and still less the conceptual objects in which the
scientific facts are cast. The reason for this is that the more 'basic'
materials are not used as tools in such a technical problem until they
have been made accessible through standardization, and by this
process their associated pitfalls identified and removed. The physics
which is used by the engineer, and similarly the mathematics used
by the experimental physicist, is, by the time it reaches him, a
robust tool. And as inexpert tool-users, they will restrict themselves
to just those properties of the tool which have been established as
safe and reliable for that sort of function.
By analogy, we can say that the scientist's handling of his con-
ceptual objects is restricted to their more gross properties, which
can be manipulated safely. It is only in the early stages of work with
deeply original concepts that their essential obscurities present pit-
falls to the scientific investigation. These usually do not survive into
the public record, for the problems are not adequately solved until
ways around these pitfalls have been charted. Examples of this
phenomenon are therefore quite rare; one of the more accessible is
Galileo's discussion of the paradoxes of accelerated motion, which
appeared in his Dialogue after some three decades of unsuccessful
 The Special Character of Scientific Knowledge 21 9

grappling with the concept. Fortunately, he achieved a practical

control over the concept of instantaneous velocity in time for the
publication of his Discourses, dealing with the sciences of mechanics;
but even then his followers and admirers found the concepts very
difficult to grasp; Gassendi, in particular, defended Galileo's
mechanics for some years before he realized that he had quite
misunderstood it. IZ The pitfalls in the handling of the objects of the
differential calculus can easily be found in the early notes of Leibniz,
and some remained to be resolved by his followers. 13 Even in the
works of Euler, some fifty years after the calculus was put into a
more or less standard form, there are few examples of the effects of
the underlying obscurity;14 but in his hands the more elementary
parts of the calculus at least, were made into the robust, general-
purpose tool which it has remained ever since.
Indeed, we can say that the neglect of the obscurities in its objects
is one of the necessary conditions for the maturing of a discipline.
So long as scientists in a field find themselves forced to engage in
philosophical inquiries to locate the causes of the pitfalls they
encounter, and are unable to manipulate their conceptual objects
12 In the First Day of the IA./ope Galileo tried to establish the paradoxical property
of naturally accelerated motion, that 'a body passes through all the infinite degrees of
slowness before reaching any assigned degree of speed.' For this he used the thought-
experiment of motion along a plane of arbitrarily slight inclination. But this brought
him to a more serious paradox: while we know, from the principle that bodies descen-
ding through equal heights attain equal speeds, that an] two paths are traversed
equally quickly, we have obvious phenomenon that the incline is traversed more slowly
than a perpendicular path of fall. GaIileo's struggles with this paradox were not
immediately successful. See Opert Vol. 7, pp. 42-53; Drake edition of DiIIlope,
pp. 21-28. For Gassendi, see]. T. Clark, 'Pierre Gassendi and the Physics of Gali1eo',
Isis, 54 (1963), 35 1-'70. In calculating the distances ttavenred through successive 'units'
of time, Gassendi encountered a pitfall by working with velocities constant through each
unit, so that the simple superposition oftbose velocities gave a sequence of distances
represented by the series, I, 3, 6, 10 rather than the square series I, 4t 9, 16. To 'save'
Galileo's law he needed to invoke an extra force; and only in 1645 did he sort out the
difficulty.
13 Tile E./y Mat1Jnlllltilai MtmlUtripts of Leilmiz, translated by]. M. awd (Open
Court, Chicago and London, 1920), although no longer accepted in all details of its
interpretation, gives a vivid impression of Leibniz's explorations, and his difficulties
with the operator tl; see pp. 90-103.
14 See the 'Remarks on Mr. Euler's Treatise Entitled Mechanics' in Benjamin Robins,
Mat1Jnllllti,fIl Trillts, ii (1751), 197-221. Robins was one of the most capable of British
mathematicians of his period; his oontributions to the controversy over Berkeley's
'Analyst' were intelligent, and his treatise on gunnery was translated into German by
Euler himself. For an illuminating discussion of the obscurities in mechanics and the
calculus during the first half of the eighteenth century, see T. L. HmkiDs, Jea"
D'.A1nnJJert, Sdmee ,,,,,1 tile Enli,"tmmmt (Oxford, 1970), cbs. 7-10.
220 The Achievement ofScientific Knowledge
reliably in the solution of scientific problems, the field cannot make
progress towards the creation of facts and the achievement of
scientific knowledge. And since these obscurities present problems
which are insoluble, and also unsettling to all but a very few stu-
dents, it is natural and indeed necessary to protect the great majority
of recruits from being troubled by them. In this way, they will do
their work of investigating limited problems, and contributing to
the advance ofthe positive knowledge ofthe field, with full efficiency.
However, a policy of a complete sealing-off of the philosophical end
of the range of problems in a field can in the long run lead to the
same diseases as the insulation of the field from influences from
other sciences and technology. For a field which goes on without the
injection. of new experiences and ideas will eventually become
stagnant; and this rejuvenation will not always come from new dis-
coveries within the field. Stimulus will sometimes come from other
parts of science or technology, but reflection on the objects of in-
quiry has, in history, led to some of the most radical innovations,
and the most dramatic progress, of science itself.
The practitioners in a field will, when they encounter undeniable
obscurities in their objects ofinquiry, generally try to minimize their
significance; and it is natural and necessary for them to do so.
Hence those who invest their time and talent in exposing them will
generally be outsiders, who may even have a prior bias against the
field or its practitioners. The debate that ensues cannot possibly
be a 'scientific' one: the problems are incapable of resolution by the
arguments accepted for the field, and social and personal elements
enter in strongly. Yet such wrangles, however inconclusive in the
short run, can be of great significance for the eventual development
of the field. The classic in this polemical literature is the pair of
essays by the philosopher George Berkeley, Bishop of aoyne: 'The
Analyst', and the 'Defence of Freethinking in Mathematics'. 15 The
first of these was ostensibly addressed to an 'infidel mathematician';
and the argument had as a conclusion that anyone who could accept
the mysteries of the calculus should find nothing objectionable in
those of the Christian faith. The immediate rejoinder, 'Geometry
IS The standard modern edition of these tracts is in Tile Woris ofGtor,t Berieley,
Bis1uJp ofClDyne, eel. A. A. Luce and T. E. Jessop, iv (Thomas Nelson & Sons Ltd.,
Edinburgh and London, 1951), 64-138. The flurry of pamphlets created by Berkeley's
attack is described in F. Cajon, .A History oftile Ctm&eptiom ofLimits "rul FlllZiom in
erm Bri";n frtnll Nt7I1t. to WtHHllunue (Open Court, Chicago and London, 1919),
cha. 3 and ..
 The Special Character ofScientific Knowledge 221

no Friend to Infidelity', is a classic of its own type of incoherent

apologetic literature; it was written by James Jurin, one of the most
successful of the mediocrities who dominated English science in the
name of the recendy deceased Sir Isaac Newton. With such a well-
defined target at point-blank range, Berkeley could produce a
second attack more carefully designed, and hence even more
effectively argued, than the first. His basic point was that the cal-
culus inevitably involves manipulating with quantities which cannot
be distinguished from zero; and when one takes a quotient of two
such quantities, one is being either self-conttadictory or hopelessly
obscure. Berkeley was far from being the first to realize this; but he
was the first man of real philosophical penettation who used this
point to question the legitimacy of the whole subject. Although his
own suggested solution to the mystery was both obscure and
fallacious, it would be incorrect to think of him as merely a church-
man who conducted some amateur investigations in the founda-
tions of mathematics. For the most likely motivation for the attack
was in Berkeley's antagonism, as a metaphysician. and as a theolo-
gian, to the Newtonian philosophy of nature which was by then a
dominant orthodoxy in English thought.
As a hostile outsider, Berkeley could bring complete clarity of
thought, and of expression, to his critical analysis of the orthodoxy
of the calculus. He could distinguish neady between mastery of
technique, and understanding of principles.
What I insist on is, that the idea ofa fluxion simply considered is not at
all improved or amended by any progress, though ever so great, in the
analysis: neither are the demonstrations of the general rules of that
method at all cleared up by applying them. The reason of which is,
because in operating or calculating men do not return to contemplate the
original principles of the method, which they constandy presuppose, but
are employed in working, by notes and symbols denoting the fluxions
suposed to have been at first explained, and according to rules supposed
to have been at first demonstrated. This I say to encourage those who
are not far gone in these studies, to use intrepidly their own judgment,
without a blind or a mean deference to the best of mathematicians, who
are no more qualified than they are to judge of the simple apprehension,
or the evidence of what is delivered in the first elements of the method;
men by further and frequent use of exercise becoming only more ac-
customed to the symbols and rules, which doth not make either the fore-
going notions more clear, or the foregoing proofs more correct. 16
16 'Defence of FreethiDking in Mathematics', section 30, p. 117.
 222 The Achievement ofScientific Knowledge
Slighdy further on he gave a brief resume of the process of the
implanting of an orthodoxy in science.
Men learn the elements of science from others; and every learner hath
a deference more or less to authority, especially the young learners, few
of that kind caring to dwell long upon principles, but inclining rather to
take them upon trust: And things early admitted by repetition become
familiar: And this familiarity at length passeth for evidence. 17
The 'Analyst' controversy is also a classic in its later effects. First,
there was a flurry of polemics between defenders of Sir Isaac, who
accused each other of failing to understand the simple and obvious
principles underlying the calculus. A British mathematician of
great distinction, Colin MacLaurin, tried to answer Berkeley in
mathematical terms, showing that the problems solved by calculus
are equivalent to those solved by geometrical constructions, them-
selves capable of perfectly rigorous proof. His book was deep and
also long, but failed in its goal of justifying the manipulations them-
selves. Thenceforth for more than half a century the problem lived
on, but at the periphery of mathematical practice; and it had to
share place with other 'mysteries' involved in the calculus, as nega-
tive and complex numbers. It was not until the early nineteenth
century that the challenge raised by Berkeley led to deep new
advances in mathematics itself; this was in the work of Cauchy,
attempting to provide, once and for all, a rigorous foundation
for the calculus. With this work, there began a new chapter in
the history of mathematics, that of rigorous proof in analysis.
Subsequent developments showed two features which might
be considered paradoxical. First, the solution of problems at each
stage opened the way to the discovery of obscurities at ever
deeper levels; by the early twentieth century, rock bottom had been
touched, temporarily at least, with the exposure of paradoxes in
the notion of·'set', and in the very structure of logic. The other
feature of this history, in which both mathematics and philosophy
were revolutionized, is that the mysteries in the formal manipula-
tions of the calculus were not entirely resolved. It was of course
universally believed that Cauchy's work had in principle made the
calculus clear and rigorous; and with that assurance, no one bothered
to observe that Cauchy's attempted justification of the symbolic
manipulations of the calculus is a pastiche of borrowed ideas, none
17 ibid., scc:tiOD ~I, pp. 117-18.
 The Special Character ofScientific Knowledge 223

of them rendered with either acknowledgment or clarity.18 For the

symbolism of the calculus, as invented by Leibniz, encompasses
such deep and basic ideas of continuity and change that it would
require a philosopher of the same genius to unravel some of its
obscurities in their own terms.
This case study from mathematics shows in a particularly pure
form the dialectic of the problems of grappling with obscurities at
the foundations of a discipline; and it is only natural that the pure
case should be in mathematics. The immediate response to their
exposure is purely defensive; then there is some worry about the
problem, at the periphery of the field; eventually a deep attack on
them yields a great development of the discipline, in its objects and
methods; and through it all the obscurities remain, but now safely
relegated to the obscurity of the schoolteachers' world.

Influences ofMethods: the School and its Cycle

The account given so far of the further evolution of the materials
which may become scientific knowledge is still deficient in one
respect. For the processes of ttansformation of content, extension
of function, standardization and differentiation, do not necessarily
go on with steadily increasing cumulative success. All these develop-
ments take place in the course of the investigation of particular
problems, governed by the controlling judgements of those who
use and adapt the existing materials in their work. The influences
from these controlling judgements, and from the rest of the body
of methods, are not capable of a simple listing. But it is worthwhile
to discuss one particular systematic influence on the direction of
research, for its prime importance in the patterns of development
of scientific knowledge.
We recall from our discussion of the conclusions of problems,
that each such solution is limited and fallible. And although the
components of the research report are tested and transformed by
the processes I have described, these processes are not like simple
quality control or sifting, whereby the 'good' is preserved while the
'bad' is rejected. For this work is done in the course of the investiga-
tion of other problems; some of them are scientific, some technical,
and some practical, but each of them as limited in its own way as
18 The aitical evaluation of Cauchy's achievements and limitations has only recendy
begun; see I. Grattan-Guiness, Tile DewltJpmmt of tile FtnIIIIltItitnu of Mllt1JnNJtietU
,AtIIIlysisfrom Elder to RiemII"" (M.I.T. Press, 1970 ).
 224 The Achievement ofScientific KnoJl)ledge
those from which their materials were derived. Hence the develop-
ment of scientific knowledge is never independent of the judge-
ments which control the investigations of particular problems; and,
since we are concerned with the direction of research, the judge-
ment of value is crucial here. Since such judgements are by their
nature very imperfect, and based on an estimation of the future,
we can ~ why the advance of science necessarily takes place largely
by way of detours.
To analyse the nature of these detours, as a basis for seeing how
they can eventually reveal a correct avenue of advance, I shall use
the concept of a 'school'. This is most easily seen in terms of the
work done on a family of problems, usually descended from some
~minal ancestor-problem. The problems are related not only by
descent, but also by their objects of investigation, a battery of tools,
and a body of methods. In these, criteria of value demarcate schools
most sharply; and a well-developed school will also have a style of
work common to its members. In addition to these technical aspects,
the school will also have its social side, in the network of friendships
and loyalties built up in the course of co-operative work; and an
institutional aspect, perhaps in the place of employment of its
nucleus, or in a stable pattern of relationships with particular in-
vesting agencies, as well as some elements of the formal apparatus
of a field, such as journals and societies.
I prefer to use the old-fashioned term 'school', even though it has
connotations of discipleship, rather than 'invisible college' or, say,
'super-problem'. For the former alternative term focuses attention
on a particular aspect of the social character of some schools, and
the latter implies a degree of definition to its work which is not
always present. Not all scientific work need be done within the
framework of recognizable schools; indeed, the most original
advances tend to come from outside the existing schools. But even
in such cases,· the work of extending and consolidating such ad-
vances usually involves a technical and social situation which can
then be considered as a new school. The extent to which organized
schools dominate the work in a field of science will naturally depend
on its technical aspects and its social history; where effective attack
on the problems requires large-scale investment and co-ordinated
work, then those working outside schools, even if not few, will
certainly be lonely.
The methods governing the accomplishment of the tasks of a
 The Special Character ofScientific KnoJl)ledge 225

school bear the same relation to those ofa single problem as strategy
to tactics in military operations. They require a longer perspective
into the future, and an assessment of more complex and imponder-
able factors; and hence extra qualities of leadership are required for
the effective direction of a school. But in both cases, decisions must
be made for the concentration of resources and work in a particular
direction, to the exclusion of others. In the case of the strategy of a
school, the carrying out of a basic decision will require more time,
and will have greater consequences for the field in which it is
operating. Yet there is the same inescapable element of risk in any
such strategy; the assessments on which the decisions are based are
conditioned by the controlling judgements of value, feasibility, and
cost.
The strategy of a school will have an obvious influence on the
problems which are chosen for investigation; and there is an in-
fluence, less obvious, but perhaps even stronger, on those materials
that are kept alive in the public channel of communication. For,
apart from those results that find a niche in a handbook collection
of very standard information, the material that is chosen for in-
clusion in surveys of a field, and then in specialized and advanced
textbooks, will reflect the judgements of value of those who make
the selections. In this task, part ofthe value ofa piece ofinformation
is its capability for fitting into a coherent account of an important
topic. Stray results which did not derive from problems in the
family worked by some school will be more difficult to fit in neady,
and so will tend to be neglected. Materials which are left out ofsuch
collections are still available, in principle, to those from other fields
who might be interested in using them as information and tools.
But they require the work of retrieval for their discovery, or even
for the establishment of their existence; and the presence of a con-
venient collection, implicidy claiming comprehensiveness for its
functions, will discourage further searching. A similar effect, in the
burying of material that does not find a natural place in a coherent
account, will occur even more strongly in the selections made for
oral teaching. In this way, the advanced training of recruits is
inevitably conservative; by these omissions, their picture of the
field as a whole is restricted to those parts which have been worked
over by leading schools. 19
19 This process of exclusion of the results of apparendy marginal fields can be seen
clearly in the historical literature on lCientific disciplines, which reflects their
226 The Achievement ofScientific Knowledge
The traditions of scholarship within the history of science have
hitherto inhibited the chronicling of the rise and decline of success-
ful scientific schools; but materials are now becoming available for
two illustrative case histories. The first is French physical science
in the early nineteenth century, which provided brilliant results, an
example of a style of work, and a model of excellence for con-
temporaries and successors all over Europe. 20 It is well known that
French influence radiated outwards, first to Germany and then
elsewhere; but general histories have tended to neglect what was
happening Pi France itself after its period of greatness. It now
appears that we should speak not of one cycle but of two; the first
being 'Laplacian physics', and the second, 'anti-Laplacian physics'.
Their creative periods can be roughly dated to the intervals 1800-12
and 1815-23, respectively. The strategy of the Laplacian school was
to apply mathematical arguments to objects of inquiry derived from
an eighteenth-century Newtonian tradition, for the explanation of
old and new experimental results. Its characteristic feature was the
assumption of 'imponderable (and intangible) fluids', consisting of
particles related by short-range centrally-direeted attractive or
repulsive forces, as the agencies of heat, light, electricity and
magnetism; and it was also hoped to explain the laws of chemical
affinity on this basis. The leaders ofthe school, Laplace and Berthollet
had prestige, wealth and influence; and they were able to recruit
the most brilliant students of the Ecole Polytechnique into their
circle, centred on their adjacent residences in the suburban village
of Arcueil. The school enjoyed some striking successes, notably

aeIf-amsciousnes8 and teaching at any given time. Thus, E. Nordcnskiold, Tile History of
Bioloo (Knopf, New York, 1928; original in Swedish, 1920-4), still the classic in its
field, devotes 3 pages out of 616 to ecology (pp. 558-61).
ao The fundamental study on science in Napoleonic France is M. P. Crosland, Tile
Sotiety of ktueil (Heinemann, London, 1967). The distinction between the two
schools, of the Napoleonic and Restoration periods respectively, with the comparison
oftheir styles and fates, has derived from my reading ofa draft ofR. Fox, 'The Rejection
of Laplacian Physics; Turning-Point in the History of the Physical Sciences in France',
to appear in k,law for tM History oftM u",t Seittl&ts. Two other important studies
of the problem are J. W. Herive1, 'Aspects of French Theoretical Physics in the Nine-
teenth Century', BritisA Jowul for tM History of Samet, 3 (1966), 109-32 , and J.
Ben-David, 'The Rise and Decline of France as a Scientific Centte', Mi1JtrV(l, 8 (1970),
160-80. It is only in recent years that historiaDs have perceived that there was a decline;
general histories, either of acience or of French culture, content themselves with de-
saibing the achievements and prestige of French science during its period of greamess.
An enriched perspective on the Arcueil group itself is provided by O. Hannaway in
his review of M. P. Crosland's book, in 1m, 60 (1969), 578-81.
 The Special Character ofScientific KnoJl)lttlge 227

the discovery by Malus of polarization of light on reflection; and

its recruits made up most of the leadership of French science for
decades to come. But all of these recruits except for the first two
(poisson and Biot) eventually deserted the school, either quiedy or
in open rebellion; and (not accidentally) its period of infl~ence is
nearly coterminous with that of Napoleon. Its weaknesses can be
seen most clearly in the programme for chemistry. Although it
stimulated some physical research (refraction in gases, as an ap-
proach to the study of the short-range forces of particles), and was
the context of Gay-Lussac's discovery of the Law of Combining
Volumes of gases, on its major problem of chemical theory its
achievements were few and it was soon superseded. Worse yet, it
was unable to comprehend the new results which were transforming
chemistry: the atomic theory of Dalton, and electtochemistry. In
physics, the limitations of the strategy took longer to be revealed;
but the school could not successfully meet the early challenge of
Fourier in the theory of heat, and the later challenge of Fresnel in
the theory of diffraction. On the social side, the style of the school
was authoritarian in the extreme; competitors and mere outsiders
were ignored or hindered by administrative means, and recruits
were expected to show complete loyalty to the doctrines.
The anti-Laplacian revolt got underway very shordy after the
Hundred Days; and a group including Fourier, Fresnel, Arago,
Dulong, and Petit captured the positions of prestige and influence
within a short time. The results for which they are remembered
belong firmly to nineteenth-century physics, and it was under their
leadership that Paris was the Mecca of the scientific world in the
1820S. But their cycle of creativity was short-lived; the early deaths
ofFresnel and Petit deprived them ofmuch-needed talent; and those
who remained could not, or did not, carry on with creative work.
The reasons for this collapse are still obscure; two possibilities are
worth considering. One is that this grouping was not a school with a
strategy; there was no grand problem giving unity and direction to
their work. Hence when the task of consolidating the initial results
became protracted (as it always does) they would too easily lose their
courage and conviction. In this they form a striking contrast to the
hardiest Laplacians, Laplace himself and Poisson; the latter con-
tinued working in the old way for a quarter of a century after the
decline of the school, and although much of his work was sterile,
some of it was of great importance. The social situation of the
 228 The Achievement ofScientific KnoTlJlttlge
Restoration grouping may also have contributed to its decline, in
several ways. Although they did not possess a new orthodoxy ofdoc-
hine, they did inherit the social institutions of science which (apart
from an increase in paid employment through teaching and examin-
ing) were not significandy changed from those of the Ancien
Regime. These produced the traditional tendencies to competition
for place, the neglect and exclusion of outsiders who lacked a
personal patron, and the shift from a career in science to a career
based on scientific eminence. As a result of this style of social
organization, and perhaps as a result of political considerations as
well, the two most profound thinkers of the decade, Sadi Carnot
and Evariste Galois, were excluded from the official community
in one way or another, and their work ignored. Finally, the absence
of a third cycle of creativity may be explained by a youthful revul-
sion against physical science, still identified with the authoritarian
Laplacianism of the Napoleonic regime; so that in the 18308 there
was no wave of young men of promise available to inherit or seize
the positions of leadership.
Another example of a 'school' which seems to be well advanced
in its cycle is the abstract style in mathematics, frequendy called
'modem mathematics'. Although there is no single figure, and no
single grand problem, which unifies this tendency, there is a shared
style which is possessed by the members of the 'invisible colleges'
occupying the leading positions in the discipline. The origins of this
movement can be traced back to the earlier nineteenth century, in
the study of mathematical objects which became increasingly
different from straightforward generalizations of the 'quantity and
magnitude' of the ancient tradition. ConsideraMe philosophical
excitement attended the invention of 'non-Euclidean geometries',
for they touched directly on the Kantian tradition. A more gradual,
but more significant growth was in the conception of an 'abstract
group'; and it is somewhat ironical that the concept was brought to
maturity and applied by Felix Klein as the key to the unification of
geometry.21 The event which in restrospeet marks the establishment
of the abstract style was the solution by David Hilbert in 1893 of one
of the deepest problems in the theory of invariants of algebraic
forms, using entirely abstract existence arguments rather than
al For a history of group theory, including Klein'. amtribution, see H. Wussing, Die
Gmesis tles _strait", Gruppm1Jqrijfts, (V.E.B. DeutlCber Verlag der WJSSeDSCbaften,
Berlin, 1969)·
 The Special CharMter ofSeimtific KnoJl)letlge 229
manipulative constructive arguments. Although at the time this was
seen as a result TPithin invariant theory, Hilbert's pupils later con-
sidered it as the result which had killed the field. JJ In the 19308
leadership was assumed by a group of French mathematicians with
the collective name 'Bourbaki'. They embarked on a programme of
re-casting all worthwhile mathematics in an abstract, axiomatic
form; and by the 19508 most of the prestigious pure mathematics
was dominated by their style. .
The criteria of value of this school are very strongly defined; and
the tendency to derogate work which involves only nineteenth-
century analysis, or applications, soon led to protests over its ex-
clusiveness. As early as 1947, John von Neumann, then one of the
greatest living mathematicians, issued a warning that such a ten-
dency could lead to the subject becoming 'baroque'. J3 The social
situation of science in the present period has also led to tendencies
which some have considered distortions. One recent critic" has
complained that the research scene is characterized by the frantic

aa See C. S. Fisher, 'The Death of a Mathematical Theory: a Study in the Soc:ioJoIy

of Knowledge', kelaw/or History 0/Ulltl Snn.e,s, 3 (1g66), 137-59.
aJ See John von Neumann, 'The Mathematician' (1947), in Col1eaetl Woris, i
(Pergamon, 1961), 1-9. As he spoke from experience of work in many fields ofmatbe-
matics, from the most directly applied, through to the philosophical, his words are
worth quoting in full: 'I think that it is a relatively good approximation to truth-which
is much too complicated to allow anything but a p p r o _ mathematical
ideas originate in empirics, although the genealogy is sometimes long and obIcure. But,
once they are so conceived, the subject begins to live a peculiar life of its own and is
better compared to a creative one, governed by almost entirely aesthetical motivations,
than to anything else and, in particular, to an empirical science. There is, however, a
further point which, I believe, needs stressing. As a mathematical discipline travels far
from its empirical source, or· still more, if it is a second and third generation cmly in-
directly inspired by ideas coming from 'reality", it is beset with very p-ave dangers. It
becomes more and more purely aatbericizjng, more and more purely I'., lOWI'.,.
This need not be bad, if the field is surrounded by correIateclsubjects, wbich still have
closer empirical connections, or if the discipHne is under the influence of men with an
exceptionally we1l-developed. taste. But there ia • paft danger that the subject will
develop along the line of least resistance, that the stream, so far from its source, will
separate into. multitude of insignificant brancheI, and that the discipline will become
a disorganized mass of details and complexities. In other words, at a peat distance &om
its empirical source, or after much "abstract" inbreeding, a matbematicalsubject is in
danger of degeneration. At the inception the style is usually classical; when it shows
sips of becoming baroque, then the danger signal is up. It would be easy to give
examples, to trace specific evolutions into the baroque and the very high baroque, but
this, again, would be too teebnical' (p. 9).
a4 See W. G. Spohn, Jr., 'Can Mathematics be Saved?' Notiees 01 1M .A..me..
MIII1Jerutietll Sodtty, 16 (1969), 8gcr4- (I am indebted to H. J. M. Bos, of Utrecht
for this reference.)
230 The Aehievemmt ofSeimtiji& KnoJl)ledge
production of fragmented results, with no institutions for syn-
thesizing important work or sharing it with those outside a coterie
of specialists. Moreover, the concentration on abstract research, and
the shaping of teaching for the training of new recruits, has had
deleterious effects in several ways. The missionary zeal for the reform
of American high school teaching of mathematics brought forth a
carefully-worded warning signed by some seventy mathematicians,
some of them eminent men in the modem style. 25 But university
teaching is the province of each group of mathematicians; and here
the damage may be serious. For there is a tendency for students
ttained by abstract mathematicians to despise the teaching of ele-
mentary and applied mathematics as tools for scientists and en-
gineers; and either to accomplish this task in a bored and perfunctory
manner, or to abandon it to the scientists themselves. The result is
a deterioration of the mathematical competence of those who need
it as an essential tool in their work. The relinquishing of a monopoly
on the teaching of its subject within a university can have serious
consequences for a discipline, even at the crude level ofthe provision
of jobs; if mathematicians do not justify themselves by their useful-
ness, then they are at risk of being considered as practitioners of a
particularly pure and esoteric art form. All these tendencies are of
course most extreme in America, but even in England some signs
of sterility can be observed. For the 'new mathematics' curricula for
secondary schools were devised by schoolmasters, with little or no
participation by research mathematicians. Rather than attempting
to retail an axiomatic mathematics based on the idea of 'set', they
bring the new ideas of 'sttucture' down towards the common-sense
level. Although these too have been subjected to criticism,26 they
may well create a new sort of mathematical common sense, in which
'matrix' is as ordinary and ubiquitous a mathematical object as
'function'. Yet there is little sign that the university mathematics
teachers recognize the importance of this development; and the
teaching of the mathematical tools of structure is frequently left
to self-taught practitioners in other fields.
as See 'On the Mathematics Curriculum of the High School', America" Mathematical
MtmtlJly, S9 (1g62), 189-'93.
a6 For an extremely vigorous argument against the British 'new mathematics' of the
Jg6os, see J. M. Hammersley 'On the Enfeeblement of Intellectual Skills by Modem
Mathematics and Similar Soft Intellectual Trash in Schools and Universities', Bulltti"
oftAt I"stitute ofMtlthemtuics tiM its Applicat;tm, 4 (1g68), 66-8S. (I am indebted to
Dr. I Grattan-Guinncss for this reference.)
 The Special Character ofS'imtiji& Kno11Jltdge 23 1
Abstract pure mathematics still attracts recruits of great talent
and vigour, and so one cannot say that it has passed the peak of its
cycle of creativity. Some of the social problems discussed above may
be the result of the over-expansion of the community engaged on a
type of work which must be done very well indeed ifit is to be worth
doing at all: but in the institutional setting of modem science and
mathematics it is difficult in the extreme to apply a severely restric-
tive policy of recruitment through postgraduate study. The danger
is that some really new insights deriving from the stimulus of ap-
plications, will not be incorporated into 'mathematics', but rele-
gated to a special field of 'application' (as computers, statistics, or
the mathematics of organizations) and so fail to achieve the rejuven-
ating effect on mathematics which will eventually be necessary.
The work of a school, like the task of investigating a single prob-
lem, involves choices and hence exclusions; and it is only in retro-
spect, ifat all, that the correctness of these decisions can be assessed.
But the difference in time-scale and social character between the
investigation of an individual problem and the work of a school is
the cause of significant differences in their cycle of development.
As I showed earlier, a problem has a natural end, when a conclusion
can be drawn. Even if this conclusion is not as extensive or deep as
that which was hoped for at the inception of the work, it exists as a
real thing. That task is accomplished, and the scientist can then
decide whether his next task should be to improve the previous
conclusion, or to tum to something else. This can also happen in
the case of a school, when a grand ancestor-problem is finally solved
to the satisfaction of those working on it. But it is more common for
the work of a school to lack such a neat conclusion. For, even when
it is organized around such a grand problem, its work will tend to
be derivative on the insights of its founder. Although its tools may
become more refined, it will still work on objects of investigation
basically conceived by him, and use only those tools which are
appropriate to them. Only when the solution of the grand problem
can be achieved by more powerful descendant tools will the pro-
gramme come to a natural conclusion. More commonly, there will
be a steady accumulation of results, but (after a time) no significant
enrichment in the knowledge of that field. The leadership has a
technical, social, and emotional commitment to the established
methods; and so, unless it is very enlightened, recruits who bring
in fundamentally new insights will not be welcomed or absorbed.
232 The Achievement ofScientific Knowledge
In such situations, the result is that over the generations the work
of the school becomes less original, less significant, and eventu-
ally obsolete. Its detour has become a cul-de-sac. If in spite of this
it retains political power, it can then become a serious distorting in-
fluence on the progress of the field.
Two lessons can be drawn from this description of the work of
schools. The first is that the continuous increase in the number of
solved problems is not at all the same as the deepening and en-
richment of scientific knowledge. In his conttast of the progressive
practical crafts with sterile metaphysical philosophy, Francis Bacon
described two extreme patterns of development: of degeneration
from a single master in philosophy, and of steady improvement in
the arts."7 He believed that a bUe philosophy could be as progressive
as the arts; but since science lies midway between the two, it is only
natural that its pattern should be a mixture. For, although there is
an overall cumulative progress, within each school there is a coasting
on the insights of the founder, with the possibility of degeneration if
the school does not dissolve when its time is up.
We also see how it is inevitable for each school to proceed along
a path which can eventually be recognized, by outsiders at least, as
a detour. For the achievement of genuine progress in the long run,
two sorts of practical problems must be solved. First, that the results
done in each special direction can make their contribution to a com-
mon stock; and second, that a field is preserved from its members
marching up a variety of separate culs-de-sac. These problems are
not automatically solved; they require certain conditions of a tech-
nical and social character. The transfer of results, which we have
already discussed as extension of tools and standardization of facts,
requires a degree of maturity in the field, as reflected mainly in the
criterion of invariance by which results ate accepted as facts. If
each school requires only invariance under changes of problem
within its narrow limits, then they will soon be talking separate,
mutually incomprehensible languages, even if they share a stock of
common words. The demise of a school then entails the complete
21 See the Preface to the Crt" l1Ul.w.titm (Works, voL iv), p. 14- There the mechani-
cal arts are described as 'continually growing and becoming more perfect. As originally
inYCllted, they are commonly rude, clumsy, and shapeless; afterwards they acquire new
powers and more commodious arrangements and CODStrucbons....' On the other
hand, 'Philosophy and the inte11ectual sciences ••• stand like statues, worshipped and
cc1ebnted, but not moved or advanced. Nay, they sometimes flourish most in the hands
of the first author, and afterwards degenerate.'
 The Special CharlUter ofScientific Kno1lJledge 233
destruction of its store of facts as living, usable materials. Only if
the work of a school has enough links with that of others for its
facts to be capable of some translation into statements about other
objects of inquiry will its productions survive in the common stock.
Even then, there will be enormous losses, since the ttanslations
cannot be perfect, and the judgements governing the selection of
such material will be different from those governing its production.
The second practical problem, of achieving the replacement of
leading schools when they have outlived their usefulness, depends
on the social maturity of a field. It is impossible for such a process
to be a continuous and painless one; and that a field is permanendy
free of revolutions, conceptual or political, may be a sign of stag-
nation. We have mentioned this problem already in connection with
the judgement ofvalue; and we shall return to it more systematically
later on.
The Historical Character ofScientific Kno1lJledge
We can now review the processes whereby a part of scientific
knowledge comes to be. Its origins lie in the investigation of a
problem, where an assertion about certain objects of inquiry (and
those artificial objects themselves) develops and is confirmed through
interaction with the external world. The problem is solved when the
assertion can be obtained as the conclusion to an argument, in which
the reports of particular experiences serve as the foundations for
the evidence. The work as a whole is accomplished by the use of
tools, applied with tacit craft skills, and is governed by methods,
including the controlling judgements of adequacy and value. If the
result of the problem is of great novelty or depth (and it must
generally be so if it is to survive all its subsequent tests and become
a part of scientific knowledge), then the problem itself is almost
certainly one involving the creation of new conceptual objects, and
probably new methods as well. The problem-situation which gave
rise to the work was quite likely to have been one involving more
than the mere extension of existing results within its field, but was
stimulated by considerations ofa philosophical, or perhaps technical,
character. The result of the investigation is far from being a perfect
and ttue statement about the world; its argument is not formally
valid, but only adequate to its function; and its objects of inquiry
are clarified only enough to function unambiguously in the argu-
ment.
234 The Achievement ofScientific KnoJl)ltdge
On being submitted to a community, the result is tested, first
formally and then informally; and if it shows significance, stability,
and invariance under changes in its objects, it becomes accepted as
a fact. Again, for a deeply novel result, this process is likely to
involve controversy; if it challenges the programme of an existing
school, it will give rise to a debate which is ultimately methodo-
logical, and hence incapable of resolution by the methods of the
discipline itself. If the result survives this criticism, then it may
become the material around which a new school comes to be; and its
problem becomes the ancestor-problem for a lattice of descendants.
Through these, the fact becomes ttansformed into versions ever
more powerful and sophisticated; and these versions eventually
become obsolete and die when the descendant-problems themselves
lose significance and are forgotten. But while this is going on, the
fact also produces extensions and standardized versions, performing
a variety of functions in its own field and in others. Eventually a
situation is reached where the original problem and its descendants
are dead, but the fact lives on through a great variety ofstandardized
versions, thrown off at different stages of its evolution, and them-
selves undergoing constant change in response to that of their uses.
This family of versions of the facts, and of their associated objects
of discourse, will show diversity in every respect: in their form, in
the arguments whereby they can be related to experience, and in
their logical relation to other facts. They will, however, have certain
basic features in common: the assertions and their objects will be
simpler than those of their original, and frequently vulgarized out
of recognition. And the obscurities concealed in the original objects
of inquiry will almost certainly remain, unless there has been a
philosophical critique in the meantime.
Such, then, is the character of a unit of scientific knowledge. It
may appear to be a very untidy and also imperfect sort of thing to
enjoy such a status. But the processes which have operated to create
this family have at the same time eliminated a host of competitors.
The particular fact survives only because it is capable of breeding
hardy descendants, who find niches in so many areas of science that
the continuous displacement of problems and whole fields, while
modifying its members, do not destroy the integrity of the family as
a whole. And the obscurity at its foundations is only a testimony to
its depth; that which is perfectly clear is, in science, likely to be
perfectly banal.
 The Special Character ofScientific KnoJl)le.....
235
To reduce the body of scientific knowledge to such elements may
seem to dissolve its unity, and indeed to destroy its reality, leaving
only a heap of pragmatically justified tools. But the destruction is
only of certain ideal of knowledge: one which demands that for it
to be real, it must be clear, distinct, and eternal. Such knowledge of
the external world may exist in the mind of God; but it is clearly
beyond the capacity of human beings, who derive their knowledge
of the external world ultimately from particular interactions with it,
observed by means of their senses. That which we can r~gnize as
scientific knowledge has achieved its state only by surviving a long
series of ruthless selections, and of drastic changes in the meaning
of its objects. It thus contains within itself a segment of human
history; its roots in experience of nature are tough and necessarily
manifold; and it still grows through involvement with new scien-
tific problems, new results, and new functions.
All this is necessary for such knowledge to be proved as genuine;
the diversity in form and function of any piece of that knowledge
and the obscurity at its foundations, are not a denial of its reality,
but a necessary condition of its existence. The unity behind the
diverse appearances lies not in a unique but hidden structure and
meaning, but rather in their common ancestry and their continued
mutual interactions. We have an aggregate, a family, of particular
versions of a penetration into the external world. No single one of
these can achieve a perfect, timeless contact with that reality; and
each is conditioned by the history of itself and of the family as a
whole. The family of particular versions stays alive, and establishes
its genuineness, by being successfully used for a variety of ever-
changing functions, each particular use being governed by the
judgements of men. It is through this variety in the particular forms,
and their individual changes, that the inner bond between them,
and the genuineness of their aggregate as real knowledge of the
external world, is created and proved.
There can be practical objections to this definition of scientific
knowledge, as those facts which have survived the disappearance
of the problems which gave rise to them, remaining alive through a
variety of descendants. For it applies only to a very small propor-
tion of the achievements of science, and in any particular case it is
recognizable only in retrospect. Surely we use the term much more
freely in ordinary speech. But the objects of 'knowledge' of ordinary
usage are of a different character from the objects of scientific
236 The Achievnnmt of Scientific K,,011Jledge
knowledge. The things we 'know' by a nearly immediate, unre-
flective sensory experience, are particular, temporary and shallow.
Knowing that this book fell to the floor is not the same as knowing
that its falling was due to the mutual gravitation of itself and the
earth. The former is indubitable, but scientifically insignificant
(unless a report of it furnishes data for a problem); the latter is
profound, but obscure in its foundations, and ever subject to evolu-
tion. For many centuries it was known that the falling of the book
was due to its earthy element seeking its natural place; and even
though we now know that the cause is universal gravitation we are
as ignorant as was Newton of the nature of that force.
How widely we should apply the term 'knowledge' is of course a
matter of convention. If we cannot bear the paradox of accepting
that genuine knowledge may be fallible, then we must ban the term
altogether from productions ofthe human intellect. But if we extend
it to include all that which at any moment is accepted as fact, we
are left without any means of differentiating between the ephemeral
and the permanent in the achievements of science. It seems best to
restrict the term to those results which are so solid that they live on
(in the fashion I have desaibed) as long as does the framework of
reality in whose terms they are cast.
Paradoxes of Scientific K,,011Jledge
It is now possible for us to consider the various paradoxes that
have been accumulating through our discussion. I cannot solve them
by a philosophical analysis demonstrating their unreality; but in the
terms of the generalized desaiption of the development of genuine
scientific knowledge we can see how they can be resolved in practice.
Generally, the paradoxes, which revolve around the contrast between
the fallibility of the endeavour and the certainty of the result, lose
their force when we appreciate that such certainty is not achieved
by a single effort. Even the productions of the greatest genius do not
survive without immediate modification, and eventual transforma-
tion, of their objects. No single result in science can be proved to
be ttue; and indeed, most are not merely untrue (as assertions about
the external world) but are also of a very temporary usefulness and
life. Similarly, the rapid succession of problems, and the continuous
shift in the objects of inquiry, winnows out all those results which
were bound too closely to the circumstances of their production.
And the methods of the work, including the controlling judgements
 The Special Cha'Mter of Scientific KfIOJl)1etlge 237
of adequacy and value, are also tested by the fruitfulness and longe-
vity of the results whose achievement they condition. Although
interpersonally communicated, and largely tacit, these methods also
evolve and mature, through the testing (significandy less direct than
in the case of scientific results) of their results against further
experience. The personal endeavour that is necessary for worth-
while scientific work is itself a very artificial and social creation. For
the individual talent and style, and the special private knowledge,
are applied in a highly stylized fashion: to the investigation of
problems concerning a given set of intellectually constructed ob-
jects, working up the materials derived from experience in accord-
ance with established methods, and drawing conclusions within
accepted patterns of argument. Even the greatest creative work, in
which all these components may be strongly modified, must base
itself on a ttadition in which such modifications themselves are a
natural development.
That such genuine scientific knowedge, as I have defined it,
should emerge from these processes, is an entirely contingent his-
torical fact. The paucity of the fields in which such knowledge has
been achieved so far, and the brevity of the periods of human
history in which such work has been successful, show that there is
nothing automatic about it. It is a particularly sensitive social
endeavour, requiring an appropriate philosophical and social con-
text, and leaders and followers of dedication and integrity. But we
know that it can be achieved, for in fact it has been. With an appre-
ciation of the complexity and contingency of this achievement, we
are now in a better position to study some of the practical problems
involved in the continuation of this work in its new technical and
social conditions.
Lm.;ts of Scientific KnoJl)ledge
To conclude our discussion of the special character of scientific
knowledge, we shall consider some of its essential limitations. First,
even such genuine scientific knowledge is not absolute and un-
conditioned. Although it has become independent of the particular
circumstances of the achievement of the results from which it
derives, it is still bound to the cultural milieu in which it developed.
If it is to be accepted as genuine knowledge in any other culture,
the objects which it desaibes, and the criteria ofadequacy and value
which it presupposes, must be coherent with the world-picture of
 238 The Achievement ofScientific KfIOJl)ltdge
the culture to which it is offered. In the partial or complete absence
of such coherence, the material cannot pass as knowledge. In a new
environment it may be adopted as a tool for the accomplishment of
practical tasks or even for the solution of scientific problems; but in
that case the objects ofthe knowledge will soon be recast so as to make
the tool meaningful and effective to its new users. Moreover, only
some parts of what seemed a coherent body of knowledge will sur-
vive the ttansfer; and it may be that those which were considered
as the essential components will be rejected as false or meaningless
by the borrowers.
All the sciences of the ancient civilizations whose direct descend-
ants are still alive as sciences in European civilization have suffered
such a ttansformation. Perhaps the best-known example is alchemy;
but in this case the manual, spiritual and mystical aspects of the
operation were so intertwined that much of the body of written
descriptions of the work later had to be rejected as spurious or in-
decipherable. Astronomy shows a neater and in some ways more
interesting pattern. For the techniques of tabulating and predicting
the positions of the moon and planets, developed in the Mesopo-
tamian civilizations, are capable of being translated into modern
mathematical notation and even of being simulated on an electronic
computer. But the objects of those position-reckonings were, for
the original astronomers, not at all the lumps of matter that we see,
but gods; and in their comings and goings it was hoped to discern
a pattern which, through comparison with events in their visible,
earthly domain, would provide a clue to their intentions. 28 Similarly,
when the great astronomer Ptolemy took over data and techniques
from this source and married it to a Greek tradition of theoretical
astronomy, his stars, although now localized on spheres in the
heavens, were still divine agents. 29 With his astronomy he hoped to
lay the basis for a sound science of astrology, and also to discover
the mystical mathematical harmonies of the divine creation. For
many centuries the technical astronomy of Ptolemy was neglected
because of its difficulty, while astrological science was a standard
part of the knowledge of all educated men, and particularly of
al For the general world-picture of this civilization, see H. Frankfort and others,
Befort Pllilo.,,,y (penguin Boob, 1949). Some detail on the astrological theory of
Mesopotamia is given by R. Labat, 'La Mesopotamie', in Histoirt GenerQIe tles Sciences,
eel. R. Talon (P.U.F., Paris, 1957), i, 73-138.
at For a discussion ofPtolemy's astrology in relation to his astronomy, see A. Panne-
koek, .A History 0/.AstrtnlMlly, pp. 160-1.
 The Special Charll&ter ofScientific KnollJletlge 239
physicians. But in early modem times the objects of asttonomical
science were transformed; Ptolemy's astronomical masterpiece was
considered an obsolete handbook of tools, his astrology was cast out,
and no respectable European astronomer since Kepler has published
investigations into the mystical harmonies of the cosmos.
In arguing the restriction of scientific knowledge to its original
cultural context I have made strong use of the distinction between
knowledge and tools. The distinction is not absolute; and in con-
nection with this problem it requires a closer analysis. When we
consider scientific inquiry as a task, knowledge about the external
world is its ultimate purpose, while the tools are only a means for
achieving that purpose. Yet tools, even those whose function is in
the performance of practical tasks, do embody some sort of know-
ledge. And as a form of knowledge, they can be seen as superior in
several important respects to verbal assertions about those intellec-
tual objects which are hoped to represent reality. As we have just
seen, they are less bound to the culture of their origin; and they
have a correspondingly greater capacity for survival and continuous
development. Yet the accomplishment of practical tasks with a
particular tool is not at all the same as having genuine knowledge
of the principles of its operation; over history, the very successful
craft techniques by which civilization was built had explanations
which we now consider utterly false. Nor is the capability of tools
for transfer an absolute thing in itself. What we find is that those
tools which perform functions related to the more universal needs
of mankind, hence those which are rather more animal than specific-
ally human needs, can be transferred and translated. This we see
from the record of archaeology and anthropology. But those which
are more specialized to sophisti~ted material or cultural functions
are restricted to their original culture in the same fashion, if not to
the same degree, as the scientific knowledge in whose terms they
receive their explanations.
Another limitation to scientific knowledge as we conceive it in
modem European civilization can be shown by a discussion of other
possible sorts of knowledge. As a bridge between these different
sorts, we may first consider the contrast between scientific know-
ledge and personal understanding. The former is the sort of material
that Aristotle said can be demonstrated and taught; it is essentially
explicit and public. Understanding, on the other hand, is private and
largely tacit. Each person's understanding of a piece of scientific
240 The Achievnnmt ofScimtijie KnollJletlge
knowledge, or even a fact, will be peculiar to himself, depending on
his own history of involvement with the materials, and on his
special skills and tastes. One can 'know' a classic piece of scientific
knowledge from an early age, but one's understanding of it can,
and should, develop as one matures. Indeed, we may say that
scientific knowledge, or facts, can exist only when the overlap be-
tween the private understandings of the objects among the mem-
bers of the relevant community is sufficiendy great for arguments
to be communicated and univocally assessed.
Now we can raise the question of whether true understanding
of important things can be derived otherwise than through master-
ing a body of scientific knowledge. Aristode admitted one excep-
tional case, which he called 'practical wisdom', deriving from a sort
of craft experience of the complex and ever-changing situations of
practical life. He also invoked 'intuitive wisdom' to provide a
guarantee for the truth of those ultimate, unproved assertions on
which any demonsttative science must rest. But this seemed to be
more a matter of recognition of what was self-evident rather than a
qualitatively deeper understanding. 30 However, in the other great
tradition in Western thought associated with the name of Plato,
scientific knowledge, publicand demonsttable, is onlyan introduction
to the real thing. This is wisdom, to be derived from an intellectual and
spiritual contact with reality, rather than through the grosser
senses; and which in the last resort is dependent upon a measure of
direct illumination. For a deep inquiry into human knowledge,
especially as its conception has developed through the history of our
civilization (to say nothing of others), these other sorts of knowing
must be taken into consideration; but that is beyond our present
purpose.
 Part III
SOCIAL ASPECTS OF
SCIENTIFIC ACTIVITY
 INTRODUCTION

IN the previous section we gave an analysis ofthe processes whereby

scientific results are achieved and scientific knowledge can come to
be. We showed how scientific inquiry is a human activity, in the
short run imperfect and fallible, and in the long run conditioned
by social influences acting in an extended time. For the sake of
simplicity, we restricted our discussion mainly to a matured 'pure'
science and also assumed the presence ofthe institutions and social
practices necessary for the accomplishment ofthe various social tasks
involved. On the basis of the materials developed there, we can now
enrich our analysis to include the social aspects of the work and so
be in a position to analyse the social problems of science created
by the conditions of the present. and the near future. I
Our starting-point for the previous analysis ofscientific knowledge
was the consideration of scientific inquiry as a special sort of craft
work. Here we will study the social aspects of scientific inquiry by
considering this work as a special sort of socially organized activity.
The starting point of this present analysis is the distinction between
the collective goals of that work, and the private purposes of each of
the agents involved in it. For the work to be successful, there must
be a harmony, or at least an accommodation, between these two
sorts of 'ends' or final causes. It is naive in the extreme to assume
that they can, or should be, identical, even in cases where the work
demands dedication and self-sacrifice from the agents. For the
establishment of that harmony, there must be certain social mechan-
isms in constant operation; but for these to perform their functions,
it is necessary in turn for those who are involved in their operation
(both as agents and as subjects) to have attitudes appropriate to their
roles in the system. In general, as the functions to be performed
I Lest I IS an outsider be challenged by scientists for my presumption in lDIlysing
their social behaviour, I may quote Jerome B. Wiesner: 'The scientific commUDity IS •
whole constitutes an extremely complex social system which is very little understood,
least of an by the scientists themselves.' 'The Federal Role in Sc:icnce and Technology',
BII1Jnj. of'M A,MIU Seinltistl (Nowmber 1962), p. 45.
244 Social Aspects ofStimtiji& Acti'Dity
become more sophisticated, the methods of social behaviour neces-
sary for the accomplishment of the relevant social tasks are neces-
sarily more subde; and the private purposes governing the
behaviour of the individual agents must be CQrrespondingly more
enlightened. There is nothing automatic about the presence of this
correlation; it is possible for any socially organized activity
(including fields of science) to work well in one time and place, and
very badly in another. From an examination of the different social
tasks involved in scientific work, we will see the ways in which an
idealistic ethical commitment is involved in its social practice. It is
not an automatic consequence ofthe study ofthe natural world; but
it is a component of the code of behaviour ofthe leading men, which
is necessary for the health and vitality of the whole endeavour.
Its presence is not guaranteed by any institutional arrangements, but
depends on accidents of history, and on the social and cultural
environment in which scientific activity is conducted.
In this Part I will consider three basic social tasks in science: the
protection of property, the management of novelty, and quality
control. The first two of these appeared as important practical
problems in the two earlier periods of great advance in science: the
seventeenth and nineteenth centuries respectively. The last has
become urgent only in the present period. This chronological
sequence is a natural one, for each earlier problem was presented by
an increase in the scale and the complexity of the social organization
ofscience, and then receded into the background under the changed
social conditions of the later periods. The progression is also of
increased sophistication in the solution of the problems. As we shall
see, the operation ofquality control in present conditions is extremely
urgent, and requires social mechanisms and personal attitudes which
are as yet far from perfect for their functions. Z
:a This seIectioD of social tab is DOt inteDdcd to be exhaustive; other important ones
include rec:ruitment and tniniDg, decisiaa on the allocation of raources, and the
JIIIDaICIDeIlt ofthe variety ofreIatioDs with IOCiety at large. I haft chosen these puticuIar
social tab for diIcuIaion becauIe of their spec:iaI relevance to the problem of ethics.
 8
THE PROTECTION OF PROPERTY

CoNSIDERING science as an organized social activity, we can say that

the collective goal of the work is the advancement of knowledge.
This is not at all the same thing as to say that every scientist must or
does devote his labours exclusively to that end. Rather, when he is
engaged on the task of investigating problems, he will be trying to
achieve several different sorts of purposes. We saw this when we
discussed the components of the criteria of value, in terms of the
different functions that a solved scientific problem can perform.
Some of these relate to the use to which the result could,be put by
colleagues or successors; while others relate to the personal benefit
that the scientist himself would derive from the acceptance of a
result as successful. This personal benefit derives from the recogni-
tion, by the relevant members of his community, of the worth of his
work. This can be known only through those results which are
acknowledged to be the outcome of his labours and of no one else.
The research reports in which the results are contained thus embody
his intellectual property; and the social protection of that personal
property is necessary if each individual is to embark on his tasks in
confidence that he will receive the rewards appropriate to his
endeavours.
The Resear,h Report as Property
As a piece of property, the research report is a rather unusual
object. The property comes into existence only by being made
available for use by others; and a research report hoarded in secret
is almost certain to depreciate in value. Nor is there a market on
which it can be sold for cash: any part ofthe result that is capable of
being sold for its applicability to technical problems must be isolated
from the rest and protected through the entirely different system of
patents for inventions. The report which embodies the scientist-'s
g-5.K.S.P.
246 Social Aspects ofScientific Activity
intellectual property thus hangs suspended in a world of its own:
in this aspect of his work, the scientist has neither made anything of
commercial profit, nor performed services for a client. All he has
done is to provide materials for others who are doing just the same
as himself, in a common activity whose ultimate purposes, in terms
of the benefit to the lay society that supports it, are remote and
diffuse.
Yet this property is none the less real and important to those who
possess it. As a certification of the scientist's accomplishment, it can
bring immediate rewards. And as an implicit guarantee of the quality
of his future work, it brings in interest for some time after its pro-
duction. Because of this predictive and fiduciary element in the
scientist's intellectuaI property, it is impossible for it to be 'alienated'
like ordinary property, without serious damage to the system of
decision and control in science. Iffor any reason one scientist ascribes
to another a larger share in a research report than is correct, he is not
merely giving him more credit than he is entided to. He is also
falsifying an important part of the evidence on which the scientific
community assesses the potential for future work of the other man,
and thereby distorting the operation of its system of government.
The personal benefits derived from the possession ofthe scientist's
intellectual property will be partly material, in the way of promotion
and the incidental rewards of prestige, but not exclusively or even
necessarily so. Even ifthe scientist is not particularly concerned with
such vulgar things, but cares only for the recognition of his worth,
or even for nothing but confirmation by competent judges of the
quality of his work, the published research report is still his basic
evidence for a claim on these benefits, and is thus his main property.
The fact that such varied private purposes can be served by the one
social mechanism for the protection of property helps to explain why
the system of scientific publication could be so successful and stable
until very recendy.
The technique of diffusing results of recent research through
published papers of authenticated quality and certified authorship is
also a very effective means for the social task of advancing the work
of science. In principle, the materials become public with the mini-
mum of delay, are guaranteed to be of at least a minimum standard
ofquality, and can be put to use without the time-consuming process
of obtaining permission, or negotiating for rights with their owner.
All that is required for the protection of the personal property
 The Protection ofProperty 247
embodied in them is that a result which is used in another paper be
cited there. In addition to its other functions, the citation thus
represents a payment for use of the material; the author of the
original paper derives continuing credit from this evidence of the
quality of his work.
This system of publication and citation might seem to be such a
natural arrangement for the harmonizing of collective goals and
private purposes that if we did not already know of the strains it is
now suffering, we might imagine it to be the result ofan ideal social
contract drawn up at the beginning of science. I However, like any
other mode of social behaviour, it is historically conditioned in its
origins and operation, has its own characteristic defects, and requires
a mixture of formal rules and an informal etiquette for its effective
operation.
Earlier Social Mechanisms for Property
The social endeavour of the achievement of scientific knowledge
does not depend on this particular mechanism for the protection of
intellectual property; indeed, the heroic age of the establishment of
modem natural science, in the sixteenth and seventeenth centuries,
lacked this system, and worked with several makeshifts. First, we
should realize that in Europe up to the Renaissance the category of
personal property embodied in an isolated piece of new knowledge
simply did not exist. In the world oflearning, the 'scholastic' method
was of analysis and citation of authorities, so that however novel a
conclusion might appear, it could be exhibited as belonging to a
strong and genuine tradition. Similarly, such practical arts as
alchemy considered themselves as participating in an ancient
tradition; and the force of the tradition in the authentication of any
result was so strong that many genuinely original works were pub-
lished as the rediscovered texts of great masters. Even in the
humanist tradition, whose style ofscholarship was in many ways the
ancestor of our own, the establishment ofa pedigree for a result was
an important part ofthe work; and in such fields as medicine debates
over real practical problems were conducted as battles of books
between conflicting traditions.
The conception of a single demonstrated result embodying new
knowledge, and belonging as property to its author, came first in
I Such an IDUIDption seems to be implicit in N. Storer, Tile soeW S:lsI"" ofStietI&,
(Holt, Rinehart, "Wmston, 1966). He IdI up • model of 'exchange' of rewards by the
community for the 'creativity' of the scientist.
248 Social Aspects ofScientific Activity
mathematics. The priority to this field might be the result ofseveral
factors: it is possible to achieve a single, brief result which is not
encased in a ttadition of scholarship or practice; and the problems
being solved in such fields as algebra were indubitably new. Also,
mathematics was a field of learning (as distinct from a craftsman's
art or a learned profession) which was practised by free-lance
individuals, usually (but not entirely) in conjunction with a mathe-
matical art as architecture or engineering. Such men already had a
system ofintellectuaI property in their inventions and constructions,
which they turned to material account in obtaining employment from
wealthy patrons. It would be natural for the conception ofintelleetual
property to extend first to the branch of learning with which they
were naturally associated, and which was also involved in their
securing of prestige. z
The first devices for the authentication and protection of such
property were crude and of a specialized application: the statement
ofa mathematical problem, as a challenge to rivals to solve it; 3 it was
understood that no one in his senses would make such a challenge
unless he had already solved the problem himself. This system
remained in use through the seventeenth century, and was even
institutionalized in the requirements for tenure ofthe chair in mathe-
maties at the College de France in Paris, which had been endowed by
the educational reformer Ramus. 4 It was the occasion of the famous
encounter between John Bernoulli and Newton over the curve
producing the path of quickest descent for a body moving under
gravity between two points. 5 But by this later time, other methods
had been developed. The first was the anagram, in which a result
was publicly stated in an indecipherable form, sometimes pending its
a The virulence of the disputes over priority in the achievement of mathematical
raub in the sixteenth century makes it plain that such 'invcntioDs', along with those of
mecbaniaal devices, were recopized IS an important form of property. The notorious
dispute of Cardano and Fernri (del Ferro) apiDst Tartaglia is chronicled at length in
O. Ore, C.4atJ tM C__li", S,IIol. (Princeton University Press, 1953), pp. 61-107.
The famous solution to the cubic equation was only one of the points at issue.
J The cbal1cDp to • public debate (probably dacmded &om traditional university
eum;n'boas) was used uosuc:casfully by TUUg1iaj see Ore, Ope tit., pp. 99-1°5•
.. This unusual system of tenure (for which IlJUllleDts can be adYIDced) required
Robenal to coac:al the intellectual property of his methods and even some of his
rauIts. Hence he was sometimes anticipated by others, and became embroiled in
priority disputes. See C. B. Boyer, A Histor:l ofMlJlllnuti" (Wiley, 1968), p. 389.
5 The c:balIenge and problem are trIDSIated in D. E. Smith, A SOW" BtHJi i" M.,M-
fUlits (McGnw-HiU, 1929), pp. 644,-55. The iDcideDt with NewtoI1 is dacnDed in
L T. More, 1.- N"". (Scn1mcr's, New York mel London, 1934), pp. 569-71.
 The Protection ofProperty 249
full confirmation by the author. Galileo used this method, and Kepler
tried unsuccessfully to decipher his anagrammatic announcement of
the discovery ofthe non-spherical appearance ofSaturn. Other users
included Hooke, Huygens, and Newton himself. Huygens even pro-
posed it to the Royal Society as a regular procedure for the establish-
ment of priority of discovery and hence the securing of property
rights. 6
Such methods, involving as they do the concealment of the result,
accomplish little more than staking a claim to property; in themselves
they do not guarantee the authenticity of the result, nor do they con-
tribute to its diffusion except by stimulus. The task of establishing
genuine communication in 'philosophy' was first undertaken
voluntarily by a small group of men, who became known as 'intel-
Iigencers'; among these were Mersenne in France, and later Henry
Oldenburg and John Collins in England. Their task was extremely
delicate; for they communicated results which had frequently been
revealed reluctantly by their authors. They had to take care that the
results were not so completely specified that the recipient could
proceed too rapidly on exploitation, and yet not so vague that a
reproduction ofthe original result could escape the charge of plagiar-
ism. From the difficulties they had, we can see that a significant
proportion of the great 'scientists' of that age were even more con-
cerned for the protection of their intellectual property, than for an
immediate realization ofits value through the prestige resulting from
publication, to say nothing of contributing to a co-operative en-
deavour.?
6 For a full summary of the correspondence between GaliIeo and Kepler in this
period (in which Galileo sent a second anagram announcing the discovery of the phases
of Venus) seeJOlultIMl Kepler Gelllmmeite Wwie, iv (Munich, 1941),478-85. In a post-
saipt to the published version of his Cut1erian Lectures, Hooke listed some of his
inventicma, describing those yet unpublished by anagrams. Among these was that for
'", "';s sit tmsio'-Hooke's Law. See Hooke, LeetitIMS C",1eriaM (169'7), reprinted in
R. T. Gunther, E.,.ly Seimee in Ozfortl, viii (Oxford, 1931), 151. Newton'. anagram,
describing his 'secret' of the method of fluDons, was sent in • letter to Oldenburg, 24
October 1676; see Tile CorrelptJtulmee of III1M Nn»ttm, eel. H. W. Turnbull, vol. ii
(Cambridge University Press, 1960); text on p. 129, translation on p. 148, decipherment
on p. 159. Huygens's suggestion was sent with an anagram of his own, in a letter to
Oldenburg, 27 January 1668-g; text and Engtish trIDSIation of the letter are printed in
Tile Corresptnulnlee of Herry OlilnUnlr" eeL A. R. and M. B. HaD, v (University of
WISCODSin Press, 1968), 360-3.
7 The properties of the cycloid were a fruitfullOurc::e of priority disputes for twenty
years; see C. B. Boyer, A History of MIJI1Ieruti&1, pp. 389-390 and p. 400. The
Newton-Leibniz dispute was a scandal whose effects damaged British matherMtics
for a run century; see More, 11""& Nn»ttm, ch. XV.
250 So""1 Aspects ofStimtijie Activity
By the end of the seventeenth century, the role of the intelligen-
ccrs was merging into that of editors of journals; although these
were not yet the specialized journals comprehensible only to the
practitioners in a field. Through the eighteenth century, original
work in natural philosophy and natural history was dominated by
gendeman-amateurs, and their approach to publication was more
leisurely and casual. A long book or a lengthy memoir would sum up
years of private inquiry; those in a medical ttadition (as Joseph
Black) would announce results in their lectures and only much later,
ifat all, write up the material for printed publication. 8 Some, like the
eccentric Henry Cavendish, revealed their results only in letters to
friends, even while advising them to publish their own work. 9
Claim-staking through sealed notes was a recognized practice, as is
well known from the history of Lavoisier's work. 10 And the book, in
which an entire field was given a definitive structure, was (and
remained for some time) an important medium of publication ofnew
results. 11

S&imtiji& Journals and Intellectual Property

It is clear that this variety of techniques could work well only
when both the set of practitioners and the flow of results in any
field were small. When property rights are secured by claim-staking
on unspecified results, and immediate diffusion is accomplished by
discreet personal intermediaries, only a thin stream of results can be
handled without inordinate delays and serious misunderstandings
about contents and property rights. The ttansition to specialized
journals as the dominant form of publication came in the haIf-
s.
a D. L. Cardwell has shown me the significance of the change in style &om the
Iemue or memoir of the eighteenth century, to the brief papers of the nineteenth.
NatunUy, both forms were in use in both periods; it is • matter of reJatiw= importance.
• Thus, in • letter to the Yorbbire scientist, Michell, Cavendish wrote: ' ••• sorry
however that you wish to have the principle kept eecret. The surest way of securing
merit to the author is to let it be known IS IOOD IS possible and those who act otherwise
commonly find themselves forestalled by others.' But Engtish D&tura1 philosophy was
ItiU IUfIicieDdy a ~club' that priority cou1d be secured without printed publicalioD. Thus,
Cavendish .ys: c••• you ought rather to wish me to show the paper to as many of your
friends as are desirous ofrading it.' (CaftDdish to Michell, 27 May 1783). Qpoted in
Ruaell McCormmacb, 'John Henry Michell and Henry CaftDdish: Weighing the Stars',
Brilisll ]tIwuIf. tM Hist.:I ofStietl&e, 4 (1g68), 147.
10 I am indebted to R. G. A. Dolby Cor pointinl out the Iipificance of this praaice.

A Hist.:I of Stinnijie MIl Tu_elll PmtNJi&.h: tM on,.,

II For. faD history of the metbocIs of publication in this period, see David A. Kronick,
MIl Dew.".,., of tM
Stinniji& MIl Tu'-.111 hal, I66S-119O (Scarec%cnr Prell, New York, 1962).
 The Protection ofProperty 2S I
century coinciding with the industrialiution of Europe and the
changing social character ofscience. The medium was already there,
in the variety of journals for an educated lay public; and the shift to
audiences which were expert rather than dilettante could well have
resulted naturally from the opportunities created by the increasing
scale ofthe activity in different fields. The social history ofscience in
this important transitional period is yet to be written; but we can
discern, as in all other aspects of science, characteristic differences
between national styles in England, France, and Germany. In
England, the journals served an overwhelmingly amateur audience,
and would be published by independent societies or by free-lance
scientific publicists. Their contents, in degree of specialization and
in quality, would naturally vary along the scale from the most expert
to the least. In France, where the great achievements of the revolu-
tionary and Napoleonic periods were concentrated among a small
group in Paris, the worthwhile journals were few, managed by
members of the scientific elite, and sharply differentiated from
the rest. 12 Rather later, in Germany, the development of great
university-based science (along with all other branches of scholar-
ship) produced the journals serving a multiplicity of specialized
fields, each journal representing a school and publishing itS work. 13
As in other respects, it was the German system which became
dominant in the later nineteenth century all over the world of
active research in science, and which with modification we have
inherited.
In its classic form, this system of specialist journals automatically
performs an important function in the protection of intellectual
property: its authentication. Science has always had obscure journals,
12 French chemistry had the great AmuI1el tie emmie; the other natural sciences fiI1ed
the 01Jlel"NtitnU sur 111 pll:1.; and the monthly BuJJeti" tie 111 SMhI Plli1tntultilJw
took short papers by the members. But the mathematical memoirs had to wait either for
the frequendy delayed publications of the Acadanie (or Institut) or of the tmIe Poly-
technique. The lesser Parisian journals, with the provincial journals and academics, were
of a lower class.
A description of the leading French journals is given IS part of the account of insti-
tutional aspects of science in M. P. Crosland, Tile SMet;y ofArcueiJ, cbs. 3 and ., PP.147-
23 1 •
13 The late start of German scientific publication is well described in H. Scbimank,
'Ludwig Wilhelm Gilbert und die Anf"ange der "Annalen der Physik"', Swliloffi Arem'D,
47 (1963), 361 -73. Gilbert took over the journal from Oren in 1798 and edited it until
his death in 1824- By the end he had converted it to a more teeImic:al joumal, but he still
had to rely on tnnslations of foreign papers to fiB it. But soon after PogendorfF took
over the journal, it became a leading and very specialized periodical.
 252 Social Aspects ofScientific Activity
in which boring or mediocre work could find a printed page. 14 But
these tended to be recognizably different from the leading specialist
journals: either provincial or colonial, publishing studies by local
enthusiasts, and appealing to an inexpert audience. For a result to
be authenticated as a contrihution to an established field, it had to be
submitted for scrutiny by one of the leaders of the field or by his
chosen referees; otherwise it would tend to be ignored by the com-
munity which constituted the field. Through the system of recog-
nized journals, the leaders of a field could effectively control the
creation of intellectual property by those who aspired to, or claimed,
membership in the field. Such a dictatorial system is clearly open
to abuse; and some of the most seminal papers of nineteenth-
century physical science were printed either privately, or in less
severe journals, because the leading journals would, or did, reject
them. Yet the system had the great merit of establishing a clear
line between the authentic and the unauthenticated in scientific re-
sults and the intellectuaI property they embodied. And it was an
essential component of the system of quality control.
It is clear that a system which operates strict controls on the
authentication of results will tend to err on the side ofconservatism;
genuinely novel results will involve their own novel criteria ofvalue
and of adequacy, and run the risk of being rejected as the work of a
crank. Occasional works of genius have been so mistreated by the
mechanism of protection of property through publication in recog-
nized journals. IS But no system can accommodate genius; and these
occasional mistakes are not peculiar to this one. However, in the
rigid ascription of property rights we can identify a defect inherent
in the system, where its performance of two functions can be in
conflict. In this case it is not only that works of deep originality and
14 I am pateful to R. G. A. Dolby for reminding me of this. It was we1l before the
war that • group of American scientists formed an 'American Institute for Useless
Research', and sugated titles of journals for the publieaticms of its members: Tile
Refuse of MtNlmt PII:lsie" Tile NIUI:lPll:lsielll]tlwul, C""PII' Ptmtlw" and Muelllift
fir ,11:1•• See letter to the editor, RftJin, of Seinltifie I1JS1n1111mts, 6 (1935), 208.
IS Authors who publilbcd important worb outside the recognized channels, from
c:boicc or DeCCIIity, include Helmholtz on the ccmsenation of energy, Y1D't Hoff on
_ , Herapath and Waterston on kinetic theory, and Newlands on the 'law of
0CIaw:a' of the chemical elements. The fint two IOOD achieved respectability, but not the
otben. Newlanda was letlOIpCCtiYe1y honoured by the Royal Society after Mendeleef
establilbcd the Periodic Table, with their Davy Medal. But he was not made a Fellow;
mel he is . , mentioned in the standard bio-bibJiolraphy of nineteenth century scientists,
1- C. P.,,1IIUrffs 1M,r.,lIistlt-liler.nstlles HMIl".,,,helt, in any of the editioDs
CODCa'DiDI the yean 1858 to 1951, vola. iii-vii, 18g8-19S9.
 The Protection ofProperty 253
genius are not accommodated by the system; the need to secure
property rights through priority of publication can aetually harm
and distort the social activity ofscientific inquiry. A result belongs to
the man who first publishes it, or whose paper first reaches the
editor of a recognized journal. In exceptional cases, the property
will be shared between two or more who have achieved the result
independently, when it can be p~oved that they could not have known
of each other's work. But the criterion of impossibility is extremely
strict; the claim to ignorance of a result recently published in a
foreign journal is not sufficient to establish independence of work. It
is easy to see the benefits of a very strict rule of priority; otherwise
there could be perpetual squabbles about the sharing of ownership
of results. Without some such formal system ofestablishment of the
ownership of a result, the adjudication of priority disputes would
depend, in part at least, on the personal memories of the parties
involved. These are always unreliable, as regards chronology
and detail; and the history of earlier disputes (the best
documented being those of Newton) shows how easily a memory
can push back the date of a discovery beyond all objective
plausibility. 16
The inherent defects in the system are not due to its particular
rules for assigning ownership of the property: they lie in the very
concept of the 'result' described in the research report, as an atomic
unit with no history of its own. If scientific inquiry consisted of
'making discoveries' or even 'testing hypotheses', one might justi-
fiably imagine there to be a very brief interval of time, at the begin-
ning of which there was ignorance, and at the end, knowledge. But a
problem, especially a deep and new one, has a complex and usually
long phase of gestation. The initial insight may flicker in and out of
plausibility, as the developing argument for it encounters evidence
which confirms or disconfirms it. The problem itself may change
in mid-course, or it may lie dormant for a while, awaiting conclusive
evidence. During all this time, the problem has no status as property:
16 D. T. Whiteside has shown that the only direct evidence for Newton'. famous
testing of the inverse-square law of gravitation in 1666 is his testimony of half a century
later. His early scientific noteboob contain DO materials which relate to such a caleuIa-
tion; but they do establish that until the early 16808 Newton was working with dynamical
models derived from a Cartesian tradition. See D. T. Whiteside 'Before the PriM';",
Jowulfor tM History of Astr01UlfllY, 1 (1970), 1-19. It should be recorded that in the
discussion after the presentation of part ofthis paper at the British Society for the History
of Science in July 1969, two American scholars cited other documents in contrldic:tion
of the thesis, but Dr. Whiteside dispatched them.
254 Social Aspects ofScimtijic Activity
it neither brings recognition to the author, nor is it protected
against appropriation.
If scientists worked in a social vacuum, this frequently lengthy
period during which a project is not property, would not be too
significant. But scientists generally need to communicate informally
through the interpersonal channel about their work; and every such
communication puts one's property at risk. In fields where the pace
of work is slow, and there is a small community, the danger is not
great: the incentive for poaching work in progress are small, and the
penalties for being caught at it are great. But in fields which are
fast-moving, large, and impersonal, one must communicate in-
formally to keep up with events, and the hazards are serious. The
protection of the property of work still in progress puts one under
pressure to engage in claim-staking publication, reporting all work
as soon as it is of passable quality. This tends to produce hasty and
slipshod work, and also discourages the investigation of those deep
and complex problems where the gestation time for the achievement
of publishable results may be dangerously long. In this way, there is
a conflict between the functions of the publication system as a means
of the protection of property, and as the means of diffusion of
authenticated worthwhile materials. This must not be thought of
as an evil peculiar to this system; for as a device designed
for the performance of several functions, it will naturally be
incapable of attaining an ideal state of excellence in all respects
simultaneously.

The Operation of the System ofProtection ofProperty

For people to operate a social mechanism, and to submit to its
operations, it is necessary for them to believe that its intellectually
constructed categories are at least in satisfactory correspondence
with the reality they experience. In the nineteenth century, the
prevailing philosophies of science assumed the simplicity of'dis-
covery'; and so the divergence between the unit of property, the
result, and the real thing of experience, the problem investigation,
was generally ignored. This simplistic view helped the system to
operate, and perhaps was necessary for its maintainance; but it
resulted in a very rudimentary understanding of the possible re-
lations between two similar scientific results. Either they were
independent, in which case the similarity was accidental, or there
was dependence, and one side was suspect of theft. The bitterness
 The Protett;", ofProperty
255
of the classic priority disputes of the nineteenth century can be
explained in terms of this simplistic picture of the process of dis-
covery: a challenge could not go very far without the ethics of the
disputants being called into question. The common practice of
the disputes being opened by friends and supporters, rather than
by the principals themselves, may have been the result of an un-
spoken etiquette adopted to contain the explosive possibilities of
such disputes. There are other explanations as well: the ruling
ideology of the selfless devotion to knowledge, combined with the
etiquette of the gendeman, who would not advance claims against
another on his own behalf. l ?
In spite of these defects, the system ofprotection ofproperty does
at least provide formal rules of authentication and an etiquette of
dispute for the assignment of ownership of work up to the point of
publication. But it provides no systematic protection for property
in the next phase of the development of a problem: the subsequent
exploitation of the result. FormalIy, any published result is avail-
able to anyone, anywhere, who wishes to use it, the only price
being its citation. But the research report may be only an interim
account of work in progress; does its author have any protec-
tion against being swamped by a rival whose superior facilities or
skill enable him to exploit the result more rapidly than himself?
If this happens, his property in the published result will rapidly
lose its value; and worse yet, the investment ofhis time and resources
in an extended research programme will be wiped out. It is clear that
the protection of exploitation rights in a result cannot be operated
through a set of formal rules. The circumstances of production and
use are so enormously various, that any fixed moratorium on the use
of results would lead to intolerable delays in some situations, and
afford no real protection in others.
Perhaps because it is incapable of formalization, the principle of
17 The most thoroughly documented study of priority conflicts, CDIDincd from a
sociological point of view, is Il. Ie. Merton, 'Priorities in Scientific Dileovery', .A.me.
SoeiolD,;elll ReTJilfl1, 22 (1957), 635-59- His ana1ysis iDdudes the CDDCqJtion of inte1-
lectual property developed here, as well as that of CODflidB between ftrious 'norms of
behaviour' in science. It is noteworthy that the putic:ipIDII in IOIDe priority disputes
would preface their claims with an aauranc:e that they disliked such matters, and thought
them irrelevant to the progress of science. See ]. w. van Spronsen, 'The Priority
Conflict between Mendeleef and Meyer, JOIInIIIl of C1wIIIieill &lilt";"" 46 (1969),
136-9. In this cue the dispute wu 0ftI' the adequacy of Mendcleev'1 aaributioll of
aedit to Meyer; and it is dear that the debate was further amfuIecl by the COIIIDIOD
conviction that there was • single true diIcovery at issue.
256 Social Aspecls ofScimtifi& Activity
fairness in exploitation of results has received no mention in the
classic literature on the sociology, or ethics, of science. Yet, interest-
ingly enough, abuses of the system of property in science through
the appropriation of exploitation rights were attempted, and de-
nounced, as soon as there was a vocationally-based community of
science in early nineteenth-century Paris. The distinguished physi-
cist ]. B. Biot, who might be called the world's first career scientist,
was involved in a number of unpleasant disputes over property.
These usually did not involve the priority of the first disco,-ery
itself; but Biot!s technique, apparendy, was to learn of a discovery
informally from a colleague or friend, and then with his experimental
skill and the superior material resources at his disposal to exploit the
discovery and report on his own series of experiments, before his
informant had the opportunity to develop the work. 11 The case of
the great mathematician Cauchy was even more notorious. On
receiving a paper for refereeing, he could not resist the temptation
of recasting the proo~ improving the result, developing and genera-
lizing it in all sorts of ways, and finally publishing it in a journal to
which he had rapid access. When the paper which had originally
stimulated him finally appeared in print, it would seem singularly
crude and poindess in comparison to the results already published
by the master.19

The EtitJuette of Citations

In the absence of a formal rule defining a method of social be-
haviour, there must be an informal etiquette which governs it. In the
case of the exploitation rights to published results, the etiquette will
depend very strongly on the particular characteristics of the field,
and (as we have seen) on the style of the local community. Even
when there is a formal rule, it must be supplemented by an informal
etiquette, since its categories can never encompass the varieties ofthe
practice it is designed to govern. This can be seen even in the
apparendy straightforward technique of citation. This has at least
two quite distinct functions. The citation of the source of materials
used in an argument implicidy places that source in the argument
.1 I am indebted to Mr. Eugene Frankel, ofPrincetOll University, for this information
onBiot.
., This and other information OIl Cauchy's style of work wiD be found in the article OIl
him by H. Freudenthal, DittitlUry ofStimli,fie BitJfrllplly(Saibner, New York, 197 1),
vol. iii, p. 1]4.
 The Prote&t;oIJ ofProperty 257
itself. There is no need to argue again for the adequacy of the
materials; that is assumed to be accomplished in the original report.
Q!tite independent of this is the function of dividing the property in
the published report, and providing an 'income' to the owner of the
property which is used, by showing that his work was fruitful.
Within each of these functions, the citation can take a multitude of
meanings. The materials themselves, the uses to which they can be
put, and the relations between an existing result and a new one, are
as complex as the history of a solved problem. The material may be
crucial, or merely incidental in the argument; it may have been
central to the first formulation of the problem, or merely a late
addition; and it may have been used as it was published or required
extensive re-working. In all these dimensions, there is a continuous
and complex scale from complete dependence to near independence.
By appropriate nuance of mention, one can under-cite without
actually stealing results, or over-cite with the effect of inflating the
value of the property of a colleague. Thus the simple system
as it stands permits a considerable range of 'sharp practice'
with scientific property, which if not controlled by an etiquette,
can be more corrosive in its eventual effects than outright
theft.
Since the citations must convey some very subde messages by a
very crude device, the etiquette of each field will impose a code for
their interpretation, whereby the entries and their possible brief
comments will convey the requisite meanings to those in the field.
Each such code will depend on the character of the problems in the
field, on the types of mutual dependence, and also on the ruling
conception of the right division of intellectual property. In every
case, it will be a purely informal, perhaps tacit and unselfconscious,
craft knowledge shared by the members of the field. Thus, in the
last resort, this aspect of the system of the protection of property
depends like the others on an informal etiquette as well as on a
formal system of rules. The relation between these two sorts of
methods of social behaviour is analogous to that between the public
and the interpersonal channels of communication. They mutually
condition each other, and are together necessary for the effective
operation of the social mechanism. The means of enforcement ofan
etiquette will in general be very different from those for formal rules;
an examination of this problem will lead naturally into the question
of the ethics of scientific activity.
 25 8 Social Aspects ofScientijie Activity
Property in lruJustrialiud Science
In the conditions of industrialized science, the effective property
of the scientist has undergone certain changes, which we discussed
earlier. For making a claim on the rewards available to him, the
scientist no longer needs to rely exclusively, or even primarily, on
formally published reports of significant results of research. Two
other sorts of evidence can now perform the same function: the
informal reports of his achievement and promise, communicated by
the interpersonal channel to those who conuol the distribution of
research funds; and formally published r~ports ofinsignificant work,
either weak in itself or hasty claim-staking. Thus two new sorts of
effective property have arisen: the informal property of personal
contacts, and the second-class property of tides of publications. The
contexts of these new forms of intellectual property are quite
different, and in general there is not much overlap between the
groups of scientists who rely on them. Nor have they entirely dis-
placed the traditional form of property; their influence depends very
much on subject and place.
The development of these new forms of property is probably the
main cause of the welcome decline of the bitter 'priority' disputes
between great scientists which so disfigured the science of earlier
periods. For neither sort is so vulnerable to appropriation of things
of great value. The decrease in sensitivity over 'priority' has another
cause, which has contributed to the change in the conception of the
unit of property itself. The common sense of the matured natural
sciences has shifted away from a beliefin the simple 'discovery', and
towards a recognition of the complexity of a problem, both in its
sttueture and in its historical development. Hence when rewards are
distributed for an achievement ofthe highest quality, they are usually
shared to general satisfaction, among those who have contributed in
different ways to the development and solution ofthe problem. Thus
the earlier common sense identification of the published research
report with the actual achievement has been displaced; and as a
result there is an awareness of the inadequacy of this simplistic
conception of intellectual property for its social functions.
As a result of these recent changes, many of the defects in the
older, rigid system ofthe protection ofproperty have been removed.
But this does not imply that in all respects the new state of affairs is
simply 'better' than the old. Indeed, if we are to cope with the effects
of such social changes, we should begin to search for the inevitable
 The Protection ofProperty 259
defects characteristic of the new system as soon as it becomes a
significant phenomenon. First, there is no guarantee that the other
functions performed by the older system will be performed so well
by the new. For example, the earlier mechanism of quality canttol
in science relied heavily on the leaders' conttol over the creation
of intellectual property, which they maintained through their con-
ttol of the recognized journals in each specialized field. With the
by-passing of this channel, both through informal contacts and
through the proliferation ofunconttolled journals, the mechanism for
quality conttol is inevitablyweakened. Asecond arises in the displace-
ment of formal rules by an informal etiquette for the operation of a
social mechanism. While formal rules have the characteristic defect of
rigidity and conservatism, an informal etiquette is more vulnerable
to a gradual distortion, in response to an aggregate of individual
pressures for the achievement of private purposes. In this case,
there is no institutional inertia for the preservation of methods of
behaviour which conttibute to the fulfilment ofcollective goals at the
expense of the achievement of private purposes. Of course, if a set
of formal rules is rejected because it is seen by the community to be
hampering its collective work, then a new set of rules, with the
associated new etiquette, can be a genuine improvement. But when,
as in the present period ofscience, the change cames in an unplanned
and 'natural' way, it is more likely that the situation is one ofa short-
term victory of private purposes over collective goals.
 9
THE MANAGEMENT OF NOVEL TY

THE achievement of new results is the goal of the task of scientific

research; and in this respect science has characteristics, and pro-
blems, peculiar to itself and to only a few other spheres of human
activity. The uniqueness of science in this respect should not be
exaggerated. Any sort of work which rises above a mechanical
routine involves the solution of new problems as times and circum-
stances change; and in matured fields of science, extending the
borders of the known can be done in a very straightforward and
predictable fashion. But in science, the achievement and management
of novelty do present problems which are nearly unique. In other
sorts of work, an organization can be quite successful even if it
merely keeps pace with gradual changes in circumstances, and does
its old job well; but in science, a field in that condition eventually
comes to be judged as stagnant. Also, in science, the achievement of
novelty of any significant sort involves a challenge to some existing
intellectual property; and the task of the management of novelty
involves the orderly destruction of personal property, for the
fulfilment of the collective goals of the work.
The management of novelty emerged as a problem distinct from
the protection of personal property, when in its scale and social
organization science developed beyond a handful of isolated, in-
dividual practitioners. When one can speak of 'schools', groups of
individuals whose intellectual property is bound up with a long-term
programme of research, and whose leaders have some control over
the authentication of results and the distribution of rewards, then
the challenge of a novel result raises particular social problems. The
resolution of these problems is particularly difficult to accomplish
by methods of social behaviour involving formal rules and routines;
but the price of a too successful defence of existing property, is the
stagnation and eventual decline of the field.
 The Management ofNovelty 261

The Destructiveness ofNovelty

As we have seen, every solution of a scientific problem produces
some change in the objects of inquiry; for the objects exist only as
the collection of their known properties. These may be explicit
assertions about details, or they may be techniques involved in
establishing conclusions about objects, or again they may be the
the more informal personal craft knowledge of subde features of the
behaviour of the objects in the course of their study. Each time a
problem is solved, the result necessarily includes new properties of
the objects of inquiry, and thereby the objects themselves have
changed. Ofcourse, in the majority ofcases, the change is slight and
not significant; a refinement of the specifications of known proper-
ties only rarely produces such surprises as to suggest the existence of
hitherto unsuspected properties. (Among the exceptions to this rule,
the most famous in recent times is the discovery of numerical
regularities in the wave-length of the spectral lines of simple ele-
ments.) Even when the change is significant, it will not necessarily
present practical problems for its acceptance and assimilation. For
so long as it does not challenge the guiding strategy of the school
that conttols the work it can be welcomed as a further contribution
to the cumulative progress of the field.
But genuine progress in science depends on the occasional in-
jection of deep novelty, and then such a challenge is made. This
need not extend to an assertion of the incorrectness of accepted
results; it is sufficient that it should claim that a new set.ofproblems,
with their own objects and methods, are a more fruitful path of
advance than those hitherto studied. Such a challenge, if successful,
is destructive; for the intellectual property of the existing schools is
thereby rendered obsolete and sterile. From the nature of the situa-
tion, such challenges do not develop gradually, by insensible degrees.
Although the roots ofthe new approach may lie deep in the past, and
its preparation requires time, there is necessarily a brief period
during which the challenge is made, recognized, and fought over.
If the challenge is successful, then there is a radical transformation
in the conduct of the inquiry, and one which may be too sudden in
its coming for those with established ways of working to adapt to
themselves. These two features, depth and destructive suddenness,
are generally accepted as defining a 'revolution', and so it is in such
cases that, following Kuhn, we can speak of a 'scientific revolution'.
The term must be used with care, for revolutions in science can
 262 Social Aspects ofScimtifi& Activity
vary in their extent and depth, as much as the revolutions in society
which range from 'palace revolutions''between factions of an elite,
to those of Russia or China.
'Scimtifi& Revolutions' in History and Philosophy
One of the most interesting things about such scientific revolu-
tions, as an historical phenomenon, is that their existence was not
recognized in the self-consciousness of science in the nineteenth
century. The experience of science during the academic period was
not so much of revolutions, where a previously effective established
system is overthrown, as of pioneering. For in most fields outside
Newtonian mechanics, the heritage of the eighteenth century was of
fields in a rather rudimentary state both in their materials and in
their social organization. The achievements of the nineteenth-
century pioneers could be seen as creating new and fully effective
scientific disciplines where none had existed before. Where this work
involved the refutation .of previously accepted theories and systems,
it was relatively easy to show that in each case they were patendy
false and misleading, at least in the light ofthe new discoveries ofthe
great pioneers ofthe field. Folk-history was an important part of the
self-consciousness of academic science; and for each field there was
developed a set of legends, in which the folk-heroes vanquished the
the false and reactionary old theories. Evidence for such histories was
easily obtained from the polemical publications of the side which
eventually won, and from recollections and eulogies of the founding
fathers. Such fields as chemistry and physiology, with the bad old
theories of phlogiston and vitalism, were natural cases for such a
folk-history; and the approach was extended back to such ancestral
figures as Galileo and Copernicus. 1
The varieties of this folk-history ofscience deserve a careful study
which they have not yet received; two features of it are of signifi-
cance for our present discussion. First, where folk-heroes did not
actually exist, they had to be invented; thus Benjamin Thomson,
Count von Rumford, was rettospeetively canonized as the author
I J. Z. Fullmer, in 'Davy's Biographers: Notes on Scientific Biography', S,imct,
155 (20 January 196'7), 285-91, discusses the influence of prevailing conceptions of
'science' and 'the scientist' on the framing ofa biography. The biographical tradition for
Davy was a fairly sophisticated one; for real hagiography one can tum to the Newtonian
biognphies. Only with de Morgan's aitica1 studies of the 18408 did the warts begin to
show. See A. de Morpn, EssiIys MIllie Lift tmd Wori ofNntJtMl (Open Court, Chicago
and London, 19 14).
 The Management of Novelty 263
of the 'crucial experiment' which destroyed the theory of a heat as a
material substance. Z Second, it could never be accepted, in the folk-
history, that a refuted theory could have been a good one, scientifi-
cally justifiable in its own times. For this would have entailed the
possibility of some serious errors in science against which integrity
and discipline offered no certain protection. And the admission of
the possibility of honest error would have undermined the propa-
ganda, and the ideology, on which the science of that period de-
pended so heavily.
It was in this context that the work of Einstein was seen as
revolutionary in its philosophical implications, for a particular
generation of scientists and philosophers. It seemed that his theory
of relativity exhibited an error at the foundations of Newtonian
mechanics, on the question of the existence of absolute space and
time. And although there had already been philosophical criticism of
that very idea (starting with Berkeley and renewed by Mach), this
successful scientific attack at what was commonly conceived as a
system of ideas that had been unchanging and apparendy true for
over two centuries had a profound effect. With the assimilation of
relativity and quantum theory, physics was conscious of itself as
being in a revolutionary state, and as dealing with obscure and
difficult objects. This condition was contrasted with that of a
previously stable, and comprehensible field; and the fact that
physics, and mechanics, were actually in the same unstable state
during the whole of the nineteenth century as they were during the
twentieth, shows how suong was the influence of the dominant
ideology on the self-consciousness of science. 3
2 The Rumford myth is perpetuated even in recent and scholarly histories of science.
Thus, in A. Wolf, .A History of Scinlee, Teelmolov tmd PlJilDsoplly ill the XVIIIt"
Cm",y, 2nd ed., revised by D. McKie (Allen & Unwin, London, 1952), we have
'Rumford's experiments clearly disposed of the caloric theory of heat in favour of a
mechanical interpretation. His scientific contemponries, however, had a sufficient
measure of intellectual inertia to resist the DeW idea. So the caloric theory survived until
the middle of the 19th century' (p. 19B). Rumford's cannon-boring experiment of
17gB was recognized as one ofthe many peculiar phenomena of heat; and in the following
two decades caloric theories of heat were used with great sophistication and success by
Laplace and his school, particularly in the explanation of the velocity of sound in
terms of adiabatic compression of air. For an authoritative discussion of the whole
issue, see R. Fox, The Calorie Theory ofGlUIsfrfJltj LiI'OtIisier to Rtf'IIIIIIll (Carendon
Press, Oxford, 1971).
3 The writings ofPopper show this sense of the 'Einsteinian revolution' in science and
in the philosophy of science very dearly. In his own apaicnce EiDstein's theory
represented (among other things) the overthrow of a dogma in science that wu the
paradigm case ofabsolute scientific truth. See Ie. R. Popper, CMljea",el au RefllllllitJIII
264 Social Aspects ofScientijie Al:tivity
It was in such circumstances that the problem of novelty became
an important one in the philosophy of science. From the time of the
assimilation of Einstein's classic results, the image of the simply
cumulative character of scientific knowledge could no longer be
maintained; it was conttary to the new common sense of the leading
field. But the question remained: how could the development of
science be ttuly progressive, if deep new advances required the
destruction of previously accepted theories? Without some clear
criterion whereby the new theory can be proved to be superior to the
old, there seems to be every possibility that the change will be for
the worse rather than for the better. Also, if the old theory is simply
rejected as false, and the achievements of the older generation are
seen vitiated by error, then those of the present one are likely to be
similarly exposed, and the development of science seems to proceed
less in the forward direction, than sideways like a crab.
These two philosophical problems, framed as the choice between
competing theories, and the relation ofthe old theory to the new one,
have seemed for some time to be crucial to the understanding of
science. For unless there are some implicit rules for the decisions
made by scientists, and some logical relation of the old theories to
the new for the assimilation of the earlier material, the genuine pro-
gress of science would be merely a chance occurrence. A variety of
suggestions have been offered by philosophers of science for the
solution ofthese problems; but none ofthem have been able to smtd
up for very long against criticisms on the grounds of internal con-
sistency and of agreement with accepted historical evidence. The
reason for this is that the actual processes of decision and assimila-
tion are too complex to be plausibly reduced to simple models; and
that while genuine progress does not depend on mere chance, it
does depend on judgements of men for which no simple and certain
rules can be laid down. 4
(IloutJedae, London, 1963), PP. 192 -3; and also 'Science: Problems, Aims, Responsi-
bilities' (preIented at the Forty-«VeDth Annual Meeting of the Federation of American
Societies for EsperimeDtal Biology, Atlantic City, 17 AprilIg63), F,.lItion Proe" • •"
22 (1g63), g61-'72-
4 The very fruitful concept of 'paradigm' introduced by Kuhn in his Stnltt"" of
Seinltifie R,.".1IIIitnu leads directly to these problems, which are not solved there. For a
suney of the subsequent discussion, see Dudley Sbapere, 'Meaning and Scientific
Cwlp', in Mirul .rul Co,.,s: EssiIys ill Cont"",...y Scime, .rul Philosoplly (Univer-
sity ofPitbburg Ieries in the Philolophy of Science, 3 (1g66), 41-8S). The most recent
work ofLabtos, in which the decision-problem is discussed in terms of the tendency of.
'research pI'OII'IIDIDe' to be involved in. 'problenHhift' that is either 'progressive' or
 The Management of Novelty 265
Even the problem ofthe assimilation ofold materials, which seems
to concern meaning rather than human action, defies reduction to a
formal analysis. For the objects of inquiry are related to experience
through a complex structure of argument, formed by the composi-
tion of many conclusions of separate problems, each with its own
history up the point of present time. A new, challenging result intto-
duces an alternative sttucture, overlapping or identical with the old
one at some points, but at others radically different or simply not
developed. Some facts from the old structure will be given an at-
tempted re-interpretation, but others will be dismissed as irrelevant
or ignored as anomalous. An 'inclusion' of all the materials of the
old structure in the new is rarely possible; but neither is it possible
for either structure to be so tight, that a conclusive refutation ofone
by the other is feasible. Hence the relation between the two struc-
tures will generally include elements of agreement, refutation and
sheer incomprehension, all mixed together.
The problems of assimilation and choice are thus insoluble at the
conceptual level; simple and general descriptions and prescriptions
cannot comprehend the complexity of what necessarily occurs. The
problems can be resolved at the practical level; but the terms of the
solution are then social and ethical rather than philosophical. The
impossibility of the philosophical solution of these problems, even
to the extent of initial plausibility, reflects the historical fact that
progress in science is not generally uniform and linear; each in-
novation in science is a huMdous enterprise for everyone con-
cerned. And on reflection it is only natural that it should be so;
otherwise scientists, and science as a whole, would be able to reap
great benefits while encountering small risks, and thus would escape
the human condition.
The Problem ofChoice: Strategies and Schools
The problem of choice is a far deeper one than choosing which of
two theories is incorrect. For the challenge posed by an important
novel result does not only relate to an existing set of materials, a
complex collection ofresults, objects and methods; it also involves a
gamble on the unknowable future, in the assessment of the fruitful-
ness of the new result in its possible exploitation and further
'degenerate', provides an illuminating formalization of the fallible judgements made in
such situations. See 'Falsification and the Methodology of Scientific :Raearc:h Pr0-
grammes', in I. Labtol and A. Musgnve, en,""", MIll 1M ero.t1l of x..w"
(Cambridge University Press, 1970).
 266 Social Aspects ofScimtiji& Activity
development. In brie~ the decision is between strategies; between
the established one which producell the results being challenged,
and an embryonic one, implicit.in the new result (or explicit in an
accompanying manifesto). Thus the problem of the management of
novelty in science is part ofthe practical problem ofreplacing schools
when they have outlived their usefulness.
A great deal is at stake in the decision on the future ofa school, and
the evidence provided by any challenging new result can never be
conclusive. We recall that a problem under investigation grows in
interaction with its materials, and that when it is completed it is
necessarily rough-hewn. This will generally be even more so, in the
case of deeply novel results; for the conquering of new pitfalls, the
forging ofnew tools, and the establishment ofnew objects ofinquiry,
require great talent, daring, and ruthlessness, and also a complete
identification of the scientist with his result. He will inevitably
develop his own criteria of adequacy for the problem, which will
diJrer, at points likely to be crucial, from those of the established
school. If every anomaly in experience, and every ambiguity in
concept, were completely ironed out before the work was presented
to the public, nothing new would ever appear. Also, genuinely new
experimental work frequendy involves using tools at, or beyond, their
limits of reliability, especially when they are used by unttained or
unsympathetic workers. 5 Hence the data itself must frequendy be
supported by a certain measure of faith, if they are to serve as a
foundation for the crucial evidence. The newly created objects of
inquiry will be designed for their function in an argument relating
to these particular experiences rather than some others. It is im-
possible for the author of a new and revolutionary result to include
in his argument answers to all objections which might come from
5 The classic cue of such difIicuIties with a new instrument is that of Ga1iIeo's tele-
soope. The existing 'spyglasses' that stimulated his invention were incapable of improv-
ing on the naked eye (or astronomical purposes; and Ga1iIeo needed to control the
quality of the fine glass produced at Venice and also devise his own system of lens-
grinding. Even then, most of the leuses he produced were inferior; and for the tint
aucial years, he simply did not have enough copies of a TCal1y good telescope to satisfy
aU the demands for demonstrations. The distinguished astronomer Magini brought a
committee of professors to look through Ga1iIeo's telescope, and it was not difficult for
them to see nothing of what he claimed was there. See Galileo Galilei, DWope on tie
Gr,.t Wor/il SystlJlU (in the Salisbury translation), ed. G. de Santillaoa (University of
Chicago Press, 1953), p. 98. For a detailed history, there are several boob by V. Ronchi;
the most recent is 11 c"flllDt1lWe iii GIIliJei , J. I&ierIu tlel1600 (Torino, 1958). Galileo's
difIicuIties with lenses are descnOed in O. Pederson, 'Sagredo's Optical Researches',
em"",", 13 (1968), 139-50.
 The Management of Novelty 267
those in the established school. For to be in a position to do this, he
would need to be in complete mastery of its techniques, and to think
in terms of its objects with full facility; and he would then be
hampered in his creative thinking about his own problem. Thus it is
possible for critics to dismiss the new work on quite straightforward
grounds of adequacy, even showing standard pitfalls which the
author has encountered. 'Brilliant but unsound' is the natural
response of honest complacency to a nascent challenge.
It is all too easy to apply historians' hindsight to the counter-
revolutionary camp, and to condemn them for obtuseness and pre-
judice against a challenge which eventually proved successsful.
Indeed, one can sometimes find cases when debate over an innova-
tion has involved the sort of personal considerations which are
rigorously excluded from the ideal ofscientific practice. But this need
not be the result of weakness or corruption of the proponents; for if
the existing criteria of adequacy for results do not give sufficient
guidance, then one must fall back on other sorts of clues to the
likely quality ofthe challenging results. These will involve the author
and his past work, and also the problem itself. If the problem has a
bad history of being the province of cranks and speculators,
then whoever tries to rehabilitate it will have to struggle against
a natural and justified prejudice. 6 Worse, if the problem has
a history of involvement in a political or professional struggle,
then any advocate of it runs the danger of being dragged down
with it. 7
Independendy of all these constraints on their perception and
assessment, those ·in an established position have every motive for
dismissing challenges to their orthodoxy. For, as we have seen, their
property is at risk. It is at this point that there is a conflict between
the collective goals of the work of science, and the private purposes
of strong groups within the community. For the advancement of a
field inevitably involves the destruction ofolder intellectual property;

6 John Ziman in PuIIlie K1UJ1IJletl", mentions the case of 'continental drift', which
'dropped into the limbo of cranky and speculative notions' for nearly a haIf-eentury
(PP·5&-')·
7 This tendency wiD be most marked in sciences which are not fully matured, and
where" the innovation is related to • folk-science hostile to an academic or 'professional'
science. The persecution of those who used 'mesmerism' (or the relief of pain during
surgical operations in the nineteenth century, is recorded by E. Boring, .A History of
Ezperimmtal Psy,1uJlogy, 2nd ed. (Appleton-Century-Crofts, New York, 1 950 ), Ch. 7,
PP·II6-33·
268 Social Aspects ofScientific Activity
and those whose property is destroyed, are hurt in the process. Two
features ofthis situation make it more serious, in some respects, than
the inherent conflict between general goals and private purposes
which is involved in the normal maintenance of quality control. For
someone who loses property through the rejection of his research
report by the journals is generally at fault for having misinterpreted
the ruling criteria of adequacy and value. It should be part of the
scientist's craft skill to know what sort of result is good enough for
authentication as genuine property. Also, in each such case, only an
individual or a very small group is involved, and their investment in
that particular problem is necessarily limited. Here, on the other
hand, reliable criteria of assessment are not furnished by the craft
knowledge of the field; and the investment that is in danger can
involve the careers of many men.
Yet the continuous advance of science requires that such conflicts
be resolved, and that the community should occasionally decree the
destruction ofthe property ofeminent members, in the service of its
collective goals. How is this to be done? If there were some simple
crucial test for deciding between competing approaches, then in-
tegrity would demand that on some occasions, an elderly scientist
would have to congratulate a young man on just having ruined his
own life's work. As an alternative, one might have an impartial com-
mittee of adjudication, taking evidence from both sides, and then
deciding which strategy is to be supported, and which suppressed.
Such a solution would have the merit of finality, but it is clearly un-
workable. The criteria involved in the assessment of competing
strategies, are so complex, subde and speculative, and the property
at stake is so significant, that for justice to be seen to be done in such
cases, there would have to be developed a system of tribunals, with
their own case-law and intellectually constructed categories of
evaluation. But since there is in principle no simple criterion for
choice in such practical problems, there can be no simple and certain
means to their resolution. In the development of a field as a whole
even more than in the investigation of individual problems, detours
and their associated waste and personal tragedy are inevitable.

Social Aspects of the Dec;s;on-Problem

Thus, for the making of decisions on strategy, and on the fate of
established schools, there is no substitute for experience, which is
gained over time. But such experience does not cumulate automa-
 The Management of Novelty 26g
tieally; and the institutions for the management of novelty can be
better or worse designed for the gaining of this experience. What is
mainly required, is for the new strategy ofwork to be given a chance.
This is not at all a matter of common sense tolerance between col-
leagues. For the leaders of the established school may, justifiable
from their point of view (and perhaps correctly) consider the chal-
lenging new work not merely as unscientific, but as anti-scientific.
Oever nonsense is far more insidious, and hence more dangerous,
than the production ofan obvious crank or dullard. By its superficial
plausibility and air of excitement, this work might easily corrupt
the minds of immature scholars and students. In some styles of
scientific work (depending on nation and field), these reflections will
be given clear voice, and then be answered in kind. 8 Personal
accusations ean easily be dragged in, under the guise of reflections
whether such inadequate work could have been done by an honest
man. The established etiquette of scientific debate serves to prevent
such sttuggles from going to their natural extreme; but when one's
property, ideology, and personal existence are threatened by what
seems to be an unfounded and unscrupulous attack (whether from a
rogue innovator or adie-hard reactionary, dependingon one's position
in the sttuggle), one will use every weapon at one's disposal.'
What seems required, in such cases, is a place for the dangerous
innovator to hide, perhaps in an obscure provincial institution,
where he can get on with attempting to consolidate his work. There
he may find a new and rival school, hurling anathemas at the older
school through the means of publication at his disposal. I have
already discussed this phenomenon in connection with personal
style. In the long run, the effective decision on the relative merits
• A minor classic in the history of scientific parodies is s. C. H. WmdJer, cUbeI' dis
Substitutionsgesetz und die Theorie des Typen' (letter to the editor), 1MIJi,s A"""""
tler Cm",;t, 33 (1840), 308-10. In this, the author reported reacting manganese acetate
with chlorine, and by successive 'substitutions' achieving a substance with the fonnula
O 2 Oa+ C1e C1e 0.+ aq. The paper was not a trick on the editor, but an unveiled
attack by Liebig on the work of Dumas. See J. P. PbiDips, 'Liebig and Kolbe, Critical
Editors', CAymiII, II (1g66), 89-97.
9 An extreme case of invective against an innovator is the review of van't Ho~ LiI
Cm",;t tltIfU l'tsIMt, by Kolbe: 'This nature-philosophy, which bad been put aside by
exact science is, at present, being dragged out by pseudoscientists from the junk-room
which harbors such failinp of the human mind, and is dressed up in modem fisbion
and rouged freshly like a whore whom one tries to smuggle into good society where she
does not belong.' Kolbe's attack was quoted by van't Hoff in his lDaupral Lecture
shortly afterwards; see J. H. van't Hoff, I""';fIIIIitm iff Seitfl&t (Springer-Verlag, 1967).
Van't Hotrwent on to achieve many honours, including a Nobel prize.
270 Social Aspects ofScientific Activity
of the rival schools and their strategies will be taken by another
generation, for whom the personal ~nfliets, and the passionate
commitment to a particular strategy, are ancient and dead history.
For the quickest arrival at this healthy state, it is best for the field to
have a loose structure, so that not every member must align himself
and his students, with one side or the other. Otherwise, orthodoxies
may harden to such an extent that the rival schools become com-
pletely self-enclosed, immune from all outside influences ofcriticism
or rejuvenation and eventually splitting the field into separate, but
equally stagnant sections. Alternatively, ifa new school does achieve
a victory, then the style of the days of bitter struggle will tend to be
canied over in its success; and all those who are not enthusiastically
with it will be counted against it, and subject to the same admini-
strative procedures.
Thus in several ways, the management of novelty in a field is
better accomplished if the structure of the field is 'polycentric',
where there are several focuses of prestige and power, not all ofthem
too sharply defined. But such a suucture is not achieved automati-
cally; its existence depends on the social and intellectual context of
scientific and scholarly work. The history ofscience has provided us,
as yet, with little detailed evidence on this problem. But there is a
striking conttast between Britain and Germany on the one hand,
and in France on the other, in the progress of the physical sciences
in the nineteenth century. The network of German universities,
each trying to build its prestige for scholarship, provided a stimu-
lating environment for men of great originality in all fields. In
Britain, the institutional structure for science and scholarship was
weak indeed, but the ttadition of independent amateur scholarship
supplemented by a sprinkling of jobs, provided the base for excel-
lent work through the century. But in France, as we have seen, the
centralized scientific establishment forced outsiders to fight every
inch ofthe way. The first generation, including Fourier and Fresnel,
were ultimately successful; but it was during the Restoration period
that Sadi Carnot and Evariste Galois worked and died in enforced
obscurity. Even later, scientists who could not secure the pattonage
of a distinguished man could find themselves in an uncomfortable
and invidious position; such was the lot of the chemists Gerhardt
and Laurent. 1o
10 Forabriefaccount on Gerhardt, see A. J.lhde, Tile Development ofModem Chemistry
(Harper Ie Row, 1964), p. 2Of. For Laurent, ibid, p. 1,6. There is quoted a complaint by
 The Mflntlgemmt of Novelty 271

Thus, it seems that the dangers that arise from the mismanagement
of novelty are not so much that bad old theories will be retained and
new ones suppressed; for in the long run, provided that there are
some independent centtes even in remote nations, good work will
be recognized. Rather it is that wherever the protection of the
established property of a school becomes too rigid, then in the area
where that school holds political control the stagnation of the field
will be felt by new and potential recruits, and the really gifted young
men who are necessary for the rejuvenation of the field will simply
not be there.
Postscript: Modem Times
Bitter resistance to novelty, like the bitter disputes over priority,
seems to have declined in the recent period. This change is doubdess
due in part to the changing common sense ofscience, with an aware-
ness of the rapid obsolescence of theories and of whole fields. Also,
the change in the location of the scientist's intellectual property,
from the monument of past achievements to the contacts for future
work, enables pastwork to be jettisoned with far less cost. In addition,
there might be a cultural influence, marked by the change from the
German master-scholar and his group of disciples to the more
mobile and egalitarian society of the American model. Because of
these influences, the practical problem ofthe management ofnovelty
is less severe, and the means to its resolution are more effective. The
weaknesses of the present situation are also easy to perceive. Not all
that is new is good, and there are dangers in the dominance of
temporary fashions on scientific inquiry. Problems which require a
long time for their maturing will be less likely to be investigated,
since their results may be considered obsolete by the time they are
achieved. The prestige given to novelty of any sort weights the
scales in favour of the younger men of brilliance, and against the
older men of wisdom. In some fields, it may be quite in order for a
scientist to become an elder statesman at thirty-five; but in others,
where years of experience are necessary before worthwhile new
Laurent: 'I have not been able to dismiss an emotion of indignation in seeing certain
chemists first caD my theory absurd, then much later when they have seen that the (acts
are in agreement with my theory better than they are with all the othen, pretend that 1
have taken some ideas of M. Dumas. Ifit fails, 1 shaD be the author, ifit succeeds another
will have proposed it. M. Dumas hu done much (or the science; his part is sufliciendy
peat that one should not snatch from me the fruit of my labors and present the offering
to him.'
272 Social Aspects ofScientific Activity
results can be achieved, the accent on youth can lead to a loss of the
inherited successful craft experience of the field. These considera-
tions are very subde, but they are relevant to the judgements by
which the leaders of fields, and of science as a whole, distribute the
rewards which control quality and set the direction of research.
 10

QUALITY CONTROL IN SCIENCE

THE social activity of science has another feature which makes it

nearly unique among all sorts of work: the social task of the main-
tenance ofthe quality ofthe products seems to be accomplished with
so little difficulty that the problem of quality control has received no
more than a passing mention in any systematic discussion ofscience. I
When one contrasts this to routine industtial production, where
quality control requires an elaborated organization, or even to the
learned professions, where tribunals stand ready to judge and punish
bad work of certain sorts, the situation in science seems nearly
Utopian. This happy state of affairs might be explained by the
moral superiority of scientists, a result either of selection and in-
docttination or perhaps even of the essential nature of the work.
Indeed, the idealistic propaganda for science, characteristic of the
academic period, did not dispel that impression. But in the present
period, we know that scientists are, after all, human; and we are
also aware of the phenomenon of 'poindess publication', better
described as shoddy science. Thus the effectiveness of th~ inherited
social mechanisms of quality control for the conditions ofthe present
and near future is open to question. Our task is to examine these
mechanisms, and then to study the conditions under which they can
operate well.
Assessments of quality are an everyday part of the scientist's
work: any material that he is considering using in any way, must be
judged for its quality in the relevant aspeets.z There are, in addition,
I The only historical study known to me which aual)'lt:l ICieDtific results by the
difFerent gndes of quality, and diIcusIes the powth-pattema of the different I0I1l of
work is K. o. May, 'Growth and Qpality of Sc:ieDtific Literature', Isis, 59 (1g68),
363-71. His sample of materials was some 1700 papers on determiDants, from the late
eighteenth century to 1920.
2 Comprehensive reviews of literature in a 6eld must provide dues to quality; 0ther-
wise the reader is faced with • IDIII of iDformatioD which is . . . . bccauIe it is undif-
ferentiated. Among the subtle devices used is the heading 'See aIIo' before the IDIII of
274 Social Aspects ofScientific Activity
several contexts in which assessments of quality are made on behalf
of the whole community; these are the ones involved in quility
control as such, and we shall concenttate our attention on them. At
the top of the scale of quality, there are the results which bear a
claim for consideration for rewards: promotion, advancement, or
special marks ofdistinction for the man who did the work. And at the
lower end of the scale, there is a cut-off point for quality, below
which a result is not given the formal authentication as property by
being published in a recognized journal. The criteria involved, and
the individuals affected, will be very different in the two extteme
cases; but we shall see that they are not totally independent of each
other, either in their workings or in their effects.
The Assessment ofQuality, from Best to Worst
The criteria on which assessments of quality are based can be
reduced to two that we have already discussed: adequacy and value.
Although these criteria can be described in general terms, their
application to any particular case involves the making ofa number of
subde, indeed tacit judgements, which depend on an intimate craft
knowledge of the work under review. It is impossible to design a
simple set ofroutine tests, by which one could assign some numerical
marks to a solved scientific problem, and then grade it on a linear
scale. Nor indeed would it be feasible to erect a formal system of
categories of quality, and train up a corps of expert assessors (such
as exist in many other spheres of activity) to operate in their frame-
work. For the techniques are so subde, the appropriate criteria of
adequacy and value so specialized, and the materials so rapidly
changing, that any fixed and formalized categories would be a blunt
and obsolete instrument as soon as it were brought into use.
The result of this special characteristic of science is that if there
are to be truly expert assessments of quality of work, they must be
made by a section ofthose who are actually engaged upon that work.
In itself, this provides no guarantee whatever that proper assess-
ments of quality will be made, or that they will be used appropriately
for the maintenance of quality. Science is not unique in this respect;
in the learned professions, the assessment of quality of work, both
the best and the worst, is in the hands of the community of qualified
practitioners. We may note, however, that the professions do have
citations too iDsignificant to be discussed in the survey; the instruction means euctly
the opposite, of coune. (lam indebted to Professor P. EDt of Utrecht for this point.)
 Quality Control in S,im&e 275
a formal apparatus of judgement and penalties for the worst work,
while in science these are lacking.
To give a rough idea of the sorts of assessments that are made, I
can distinguish several classes of scientific work, and the standards
by which they are defined. The best of all scientific work is that
which survives, through all the many testings and ttansformations,
to become genuine scientific knowledge. But this can be known only
in retrospect; and so a sober assessment of any new result will not
place it in the 'immortal' class. However, one can reasonably predict
that a given result is likely to yield 'enduring facts', which will
survive the demise of the original problem; if that problem was a
deep and difficult one, and the result is capable of development and
extension, then it is entitled to be considered as first class. Now, even
a temporary fact is no mean achievement, for its existence shows
that the solution to the original problem had depth rather greater
than the explicit statement of the conclusion would guarantee; work
which achieves this success is certainly good. Even this grade is not
the minimum quality necessary for a result to be worthwhile in the
advancement of its field. So long as it meets the appropriate stand-
ards ofadequacy, and shows its value by being put to use by others,
however briefly, in subsequent work, the research which produced
it can be considered competent.
These classes are defined in very general terms, and are based on
predictions of the future career of a result. In any field, there will be
symptoms whereby a skilled man can form a judgement of the
prospects of a particular result. These will be enormously various;
elegance of argument may be basic in one discipline but irrelevant
in another; novelty of the problem or surprise of the solution may
be crucial or insignificant; concern for fine detail may be the mark
of the master-craftsman, or of the bore. Where masses of reliable
information are required for the investigation ofnew problems, then
work which does very little more than provide such information may
be quite competent. But in more theoretical or speculative fields,
where the boldness of the problem and the depth or elegance of the
solution constitute the value of the work, then routine problem-
solving along a narrow path, with standard techniques, may have
no value other than the political one for the scientist who publishes
the work. Thus quality depends not merely on the character of the
work done but quite strongly on the field in which it is placed.
Because of the subdety and particularity of the criteria of quality,
276 Social Aspects ofScientific Activity
competent judges must be in or very close to the field, and must
also be capable of producing work of a similar degree of quality to
that which they are assessing. For this reason, the assessment of
work of the highest quality, and the social tasks depending on such
assessments, must be reserved to the aristocrats of a discipline,
rather than being assigned to a democratic procedure or to the
laymen who administer the funds invested in research.
The assessment of quality at the lower end of the scale does not
require such subde and speculative judgements. But it does involve
questions of ethics, for a published result which is not even com-
petent, has a false claim to authenticity; and the referee who passed
it for publication was either lax or incompetent. A result may be
weak in a number of ways. In experimental fields it is possible for
reasonably good data to be incapable of being processed into
evidence which functions adequately in an argument. The desired
conclusion may be a thesis more general than the evidence available;
or it may be that the argument is a retrospective, ad hoc attempt to
explain the data that has been produced. In theoretical fields, the
argument may be good in parts, but those parts may lack the neces-
sary coherence for the conclusion to follow from them. A weak
result may be capable of improvement to the point of being com-
petent or better; and in such cases the referee can offer advice. And
the borderline between weak results and those which are just
competent is necessarily indistinct.
At the lower end of the class, the weak results merge into the
vacuous ones: those for which nothing good can be said. In a
matured field, the production of vacuous results is reserved to those
scientific workers whose self-respect, if they have any, derives from
other areas of their experience than the advancement of their field.
For the body of craft knowledge of methods of a matured field has,
among its functions, enabling a conscientious worker to proceed on
limited tasks without encountering fatal pitfalls, and at least pro-
ducing data and information with some content. However, this does
require some patience, diligence and judgement. If one seizes the
first plausible set of experimental data to appear, confects an argu-
ment into which they can somehow fit, and then writes it up for
publication, one can be fairly sure of achieving vacuity. In a
theoretical field, one can stan with some objects of inquiry which
have some distant relation to experience, concentrate attention on
their mutual relations at the expense of those aspects with real con-
 Quality Control in Scim&e 277
tent, and paste together an argument, verbal or mathematical,
from which are drawn invalid conclusions about non-existent
things.
Only the expert eye can distinguish between a vacuous research
report and one with some real content. By the use, or abuse, of craft
skills, it is possible for slipshod experimental work to be dressed up
in a conventional description, for unsound data to be converted into
empty information by dubious mathematical techniques, and for an
incoherent argument to masquerade as close reasoning. 3 Moreover,
this approach to research and publication not only saves time and
energy but also pays dividends in the number of publication-points
achieved. For the first vacuous research report can spawn another, as
soon as the scientist has found an error in his data; and so on. In
theoretical fields it is even easier, since every existing cobweb of
argument invites immediate further articulation. Indeed, the
question for an analysis of quality conttol and ethics in science, is
not so much how weak and vacuous work can possibly exist, but
rather how such weeds are kept from choking the very artificial
growth of genuine scientific inquiry.4 For the present, we will
resttiet our discussion to matured fields, where the social task of
quality control has at least the technical means for its
J It is not common for a scientist to identify productious of ftCUOUS research publicly;
only a master of a field could, and should, do so. For an example of this type of critical
aualysis, see F•Yates, 'Theory and Practice in Statistics',]mIru1 oftile RoytU S,"timttU
SDtiet;y, Series A (Gmerlll), 131 (1g68), 463-7.. This was his Presidential Address to the
Royal Statistical Society. According to him, one main cause of vacuous work is the
tendency of university teaehen ofstatistics to publish 'pure' research in a subject whose
origins and strength lie in its association with pnctice. In fairness, he also cited ftCUOUI
'applied' research; the most famous being that of using aerial photographs for assessing
the damage to German production caused by the mass air nids on the towDs during
World War II. Both the bombers and their statisticiaDs conc:entrated on the town centres,
in apparent ignol"lllce ofthe fact that German industrial cities, unlike the English, have
a residential centre and industrial suburbs I (I am indebted to Mr. A. B. Royse for this
reference.)
4 Here we can see how my aualysis relates to that of Popper, on the question of the
role ofcriticism in science. He used the 'critical attitude', a willingness and commitment
to submit one's theories to test, IS a criterion of demarcation between the pnctitionen of
'science' and of 'pseudo-science', and hence between those categories themselves. (See
'Science; Conjectures and Refutations', ch. I of C",jeetwes "" RefuttltiMU (Routledp,
London, 1963).) In the years since his first iDsight, the question of the penoaal motives
ofscientists bu been seen IS nther more complex, and the deman:ation correspondingly
less neat. But we can agree with his basic insight, and re-state the dichotomy on a
practical, rather than an epistemological, basis. Then we can say that the absence of.
critical attitude among the members of a scientific community is a tllfIU of a degenera-
tion into vacuity and corruption, and also a sip of ~ domiDation now or in the near
future.
IO-S.K.S.P.
278 Social Aspects ofScientific Activity
accomplishment. 5 In immature fields, the situation is otherwise; and
the tragic problems encountered there require a separate discussion.
Comparative Examples ofQuality Control
Although quality control in science is operated with a minimum
of formality, the nature of the task is the same as in other sorts of
socially organized work. By analysing the process, particularly at the
lower end of the scale, we will discern similarities and differences
between science and the more common situations, and so be in a
position to analyse the special problems encountered in science. The
function of quality conttol is to ensure that the users of the products
can rely on their being of a certain acceptable standard. The task is
divided into several phases: establishment of criteria of quality and
setting standards in their terms; testing the set of products for an
assessment of their satisfying the standards; and enforcing the
regular adherence to the standards by a system of penalties and
rewards on the operatives. In ordinary industrial production, any or
all of these phases may become highly elaborated. The first phase
involves decisions at the highest level, since more rigorous standards
lead to higher direct and indirect costs. The task of testing has
developed into an applied science in its own right, and its practice
can involve specialized equipment and sophisticated statistical tools.
And the enforcement phase involves politics (in the wider sense),
since it is assimilated into the general system of government of a
factory.
In general, the system of quality control in industrial production
is run along the same lines as those of decision and command:
rigidly hierarchical, with the unskilled workers at the bottom of the
pyramid usually being treated as simple stimulus-response machines.
Their products are tested on a routine basis to simple standards by
5 The classification of degrees of quality given here agrees well with the criteria
implicit in the 'Kcndrew Report', RtltWt oftile Worii", CAoul on MolenJ. BioloD,
Cmnd. 3675 (H.M.S.Q., London, 1968). Giving their impression that quality is very
uneven, the authors say: 'Some work in the 6eld is of the highest class but there is much
that is concerned with problems irrelevant to the real growing points in the field; work
which may not be downright bad but which contributes little to the main advance of the
subject -which, in other words, is at the best dull and at the worst trivial Our exami-
nation of the literature leads us to think that this is in marked contrast to the posi-
tion in the U.S.A., where as well as • good share ofthe work ofthe highest class there is
• peat deal of the sound and competent, perhaps uninspired but at least relevant,
work which is essential to sustained development' (para. 9, p. 4). (I am indebted to Dr.
J. Wootton for this reference.)
 Quality Control in Science 279
another set of operatives. Rewards are for large quantity rather than
for excess of quality, and the penalties for inferior quality are
assimilated to those for insufficient quantity, and are generally of
the crudest sort. Thus, in this situation, the technical problems of
quality conttol are of a purely administtative sort. Where skilled
work is being done, the system of quality control, although the
same in its essentials, must be more subtle and refined in its opera-
tion. For it is impossible to make an adequate assessment of the
quality of a craftsman's product, by simple routine tests. In such
work, the worker is constandy making intuitive, perhaps tacit judge-
ments of the quality of each operation; if these were not necessary,
the manipulations could be performed equally well by unskilled
hands or machines. The product of the work is thus conditioned by
a multitude of judgements of its quality, relating to every phase of
its making. And from the mutually reinforcing touches of excellence
comes the quality of craftsman's work, which reveals itself only in
its performance, and cannot be categorized in a small set of measur-
able attributes. Hence the control of quality of craftsman's work
must to some extent be entrusted to the craftsmen themselves.
This can be done, in spite of the divergence between the private
purposes of the craftsmen and the goals set by the hierarchical
organization employing them; the conditions under which it is
possible will be discussed in connection with ethics.
A situation more like that in science can be found in the learned
professions. The practitioners there are not placed in a comprehen-
sive formal hierarchy, and the profession as a group is almost entirely
self-policing. The establishment of criteria of quality, the setting of
standards, and the testing of work, are all accomplished informally
under the guidance of the recognized leaders, with work of high
quality as the main focus of attention. The recognition and reward-
ing of high-quality work has important functions in the system of
indoctrination of recruits, in the maintenance of morale, and in
relations with the lay public. Those whose work is merely competent
or even weak can labour in obscurity, merely earning a living. But
a genuine profession will have formal institutions for trying cases
of very bad work, and imposing punishments, even to the point of
expulsion. These are necessary because of the special relations of
professionals to their clients, and perhaps also because ofthe dangers
of scandal caused by seriously aggrieved clients. Even here, it is not
so much simple incompetence which attraet8 such penalties, but
280 Social Aspects ofScientific Activity
rather the abuse of the client's confidence by actions serving the
private purposes ofthe professional. Accordingly, such professionals
as solicitors and dentists are more at risk of punishment, than, say,
surgeons or engineers.
The Levels ofQjullity Control in Science
In science, by conttast, the situation seems to be one of nearly
perfect happy anarchy. There is no formal hierarchy of decision and
control; and the community of science has no institutions of
punishing or expelling a wayward member. A university may revoke
a man's tide to its degtee, but in general the subject-societies (usually
misdescribed as professional associations) have no such penalties at
their disposal. Indeed, the system of quality conttol in science has
only one point where formal procedures operate: the assessment ofa
research report by the referee of a recognized journal, for authenti-
cation as a published paper. The penalties involved in this procedure
is a private affair between the referee, the editor, and the author; and
no punitive action is taken against a scientist who submits one, or
even many bad papers. But the penalty is none the less real for
being discreet. Time and resources have been invested in the
investigation of a problem, and with the rejection of the paper, no
property has been created. Work has been wasted, and there will
be an embarrassing gap in the tirne-sequence of the scientist's
publications.
This is how the matter works in principle; and in situations where
the leaders of a field can and do exercise tight control over the
authentication of results this discreet quality control at this mini-
mum level is quite effective. But as we have seen, in the present
period the mechanism does not work perfectly; and so it is necessary
to examine the ways which it is maintained to the extent that it is.
The referee for a journal, and the editor acting on his advice are, as
we have seen, the effective agents in the system of quality conttol at
the lower level. The referee's task is a particularly subde sort of
craftsmanship. He must not only understand the technical material
under review; he must also be able to reconstruct in his own mind
something of a history of the investigation, to detect errors of com-
mission and omission. As a conclusion of his inquiry, he must decide
on a recommendation for publishing, and also offer suggestions for
improvement. His criteria ofquality will depend not only on the field
in which the problem is located, but also on the particular style of
 Quality Control in Science 281

the journal and interests and competence of its audience. Although

the referee may be given general guidance by the editor on the
principles of publication of that particular journal, he will still need
to i~terpret these for the special case of each research report under
reVIew.
Because the referees cannot be segregated off from the body of
working scientists in their field, they cannot be made accountable to
a separate hierarchy of conttol. A working scientist who is acting as
referee will find himself in the invidious position of being prepared,
on occasion, to deny authentication to the work of colleagues and
friends. In order that referees can accomplish their delicate task
properly, their personal skill at this craft must be supplemented in
several ways. First, they must have a commitment to the main-
tenance of the standards of quality in that particular journal, or in
the field in general, so that they can see the damage done to the
immediate interests of their unsuccessful colleagues in the light of a
higher good. Also, they must be protected by a strict etiquette which
makes their relation to each author as impersonal as possible, so as to
minimize the risk that disagreements over assessment will become
personal disputes. And, finally, they, and their editors, must them-
selves be subject to the operations of some mechanism of quality
conttol.
This higher level of quality control must operate directly on the
journals, and through them on their editors and then on the referees.
Without this, there would be no barrier against the spreading of a
decline of standards among referees, consequent on a loss of morale
or commitment among some. The outcome of such a development
would be a Gresham's Law for scientists' intellectual property,
applying directly to work which aspires to be good or competent: if
property of equal political value, embodied in formally authenti-
cated research reports, can be purchased at a lower personal cost in
time and work by submission to a lax journal, then only the saindy
minority will penalize themselves and their families by continuing to
work to unnecessarily high standards. The prevention ofsuch a state
of affairs cannot be accomplished by formal procedures: the com-
munity of science has no institutions for withdrawing 'recognition'
from a journal which publishes inferior work. Instead, there is a
very informal ranking of journals in a field by their prestige; and
although such a list will never be published, it can be an effective
means of control over editors, referees, and authors as well.
 282 Social Aspects ofScientific Activity
The mechanism here is competition for the limited supply of
rewards available to individual scientists; and it can operate because
the assessment of the quality of a research report is partly based, in
practice if not in principle, on the journal in which it appears. A
paper appearing in a journal of high prestige has a much better
chance ofattracting attention than one in a journal of low prestige.
The 'better' journal will be more widely distributed, it will be
scanned more carefully by active scientists, and a paper there will be
given a greater initial credit for quality. Hence it is in the personal
interest of the author to publish his papers in the journals of the
highest prestige which he can reach. To give his work to a weaker
journal does not merely cost him the risk of under-valuation of the
result; it also associates him with men who can do no better. The
editor also receives rewards for his identification with a journal of
high prestige; and suffers the penalty ofrelegation to obscurity if his
journal is weak. It is then in his interest to establish a good reputation
for his journal so as to attract papers which will maintain and enhance
that reputation; conversely, if he publishes papers with which a
good man would not want to be associated, he runs the risk of
losing his high-quality contributions. 6
All this is a matter of degree, of course; in a large field, there may
be several 'leagues' of scientists and journals, distributed naturally
over ranges of quality, with the hope of promotion and the fear of
relegation between leagues functioning as the immediate rewards
and penalties in this system ofquality control. Since the aspirants for
promotion will model themselves on the style of their superiors, the
whole system is ultimately governed by the distribution of rewards
for work of the highest prestige. However, not all journals, or
scientists, are in this system. There is the mass of those who,
through geographical isolation, lack of resources, or lack of talent
or commitment, cannot hope to produce intellectual property
authenticated within the dominant system. They then settle down,
either to working within their autonomous but isolated local system
6 Evidence of an imperfection in this informal system of quality conttol on journals
arises from. report on recent decisioDs ofthe American Geophysical Union concerning
publicatioDs. A new polic:y is described IS foUows: 'A review of the quality IS we1las of
the length of papen by a IfOUP which is separate from the editon. This amounts to
post-factum refereeing of the papen and should provide guidance to the editon, to the
publicatioas planning committee and to the AG.U. CounciL' Letter to Sri,,"t, 166
(1969), 4J-4t from S. F. Singer. It would appear that a amunittee accountable to the
CouDCil and DOt to the editors will monitor the published journals and make assessments
of quality (or use in future pDdance of editors.
 Quality Control in Science 283
of prestige, or to production of second-rate property in the titles of
weak or vacuous reports in the many obscure journals that coexist
with the few significant ones. The first of these groups, of 'colonial
science', have been with us for a long time; and although it frequendy
involves tragedy for its members, in exile from 'metropolitan'
science, it does not constitute any threat. 7 But the second group, of
'shoddy science', do represent a danger; for if their characteristic
property is not kept firmly in second class status, the whole system
of quality control for genuine science could easily collapse.

Leadership a,uJ Prestige

Thus the question, 'Who controls the controllers?', has brought
us to the leaders of each field, who control the rewards of prestige
(and the penalties of obscurity) which operate on journals as well as
on individuals. At these higher levels, prestige becomes an important
component of the mechanisms of conttol. It functions as the 'cash
value' of the scientist's intellectual property as assessed by his col-
leagues; on the basis of his prestige, he can command personal
benefits and the resources necessary for the continuation and ex-
pansion of his chosen work. Also, at this higher level, the criteria of
assessment of quality are significandy different from those lower
down. Adequacy can be presupposed, so the dominant component
here is that of value; and since this depends on a judgement of the
future, it is intimately related to the task of directing current re-
search. As a result, quality control at the highest levels is inseparable
from the general tasks of leadership of the field.
The informal hierarchy of control does not end at this point;
and this is just as well, for otherwise it would be impossible to main-
tain any unity or coherence in the endeavour of science. Since there
is no class of active scientists who operate independendy of all fields,
the operation of quality control on whole fields and on their leaders
must be the task of a community of leading scientists of which they
themselves are members. The relations between fields of science
are of many sorts, and in constant change: there may be areas of
overlap of problems, they may be associated through the provision
7 For an aualysis of 'colonial science', which was the condition of American science
until well into the twentieth century, see Donald Fleming, 'Science in Australia, Canada
and the United States: Some Compantive Remarb', Prot. IotA Intmlllti01UlI C""".tU
ofHistor;y ofStitn&t, 1962 (1964), pp. 172 -eJ6; and also James R. Killian, Jr., 'Towards.
Research-Reliant Society', in Scientt liS II CullUl'II1 Fortt, eel. H. Woolf (Baltimore,
1964).
284 Social Aspects ofScientific Activity
of tools, or one field may be the parent of another. Whatever their
internal relations, they are bound together, in groupings of various
sorts, by the competition for limited resources and the need to allo-
cate them wisely. A special class of leaders, of exceptional qualities
of statemanship, is required for the proper accomplishment of this
latter task. For the judgements involved are concerned with quality
and direction of research in areas where the person cannot have any
personal craft knowledge. The men involved must have the ability
to extrapolate widely from their own scientific experience in order to
assess quality and promise from the evidence provided to them; and
they must also be able to put in proper perspective the claims for
fields close to their own research interests and those remote from
them. 8 Their decisions constitute a mechanism of quality control
on whole fields; for a field which does not measure up to its com-
petitors will be deprived ofresources and thereby be sent into decline.
This crucial set of tasks of ultimate decision and control has
traditionally been accomplished within a particular institutional
framework; in those nations where the nineteenth-century German
system was followed, the universities had the central role. For,
although a man's working colleagues might be scattered allover the
world, his career would be within the university system; and his
advancement in that career would depend in the last resort upon the
judgements of colleagues in a particular university. In the univer-
sities, too, there is competition for limited resources; but there the
decisions are taken by scholars from such a wide variety of discip-
lines that in respect of the technicalities of anyone of them they are
• Thus, looking again at the 'Kendrew Report', we see that the aiterion of quality
implied in the pbrue 'mncemed with problems irreleYant to the real growing points of
the field' is capable of more than one interpretation in practice. For although the leaden
of a field may be very sensitive to new approaches coming from young men in obscure
ccntra they may be quite convinced in their mllective wisdom that their own investi-
ptioDs, DOW maturing after some years of work, CODIIitute the true 'growing points'.
No set of formal rules will distinguish in practice between the many subtle gradations
between these two positions; hence there is DO substitute for wile leadership.
In this context, Szilard'. desaiption of the effects ofeven the best-intentioned system
of ccntralized direction of resean:h, is worth recalling. '••• the scientific workers in need
of funds would concentrate on problems which were ccmsidered promising and were
pretty certain to lead to publishable results. For a few years there might be a great
increase in scientific output; but by going after the obvious, pretty soon science would
dry out. Science would become IIOIDdhiDg like a parlour pme. Some tbinp would be
CODIidcrcd interesting, others not. There would be fuhions. Those who followed the
fuhion would set pants. Those who wouldn't would not, and pretty soon they would
learn to follow the fabioo too.' Qpoted from 'The Mark Gable Foundation', in The
ytlil, ,llhe Dolllli"" p. 101.
 !Luality Control in Science 285
but educated laymen. Hence, within the universities, the assessment
ofa man and of his work, reached either through informal consensus
or by formal deliberation, will have a very different basis from that
within a field of science. On the one hand, colleagues will be able to
judge him from their personal experience of his work outside his
speciality; and this certainly gives an indication of the capacity for
leadership. But for an assessment ofthe man and ofhis field (the two
are closely bound together in the case of a leading scientist), his lay
colleagues must rely on evidence ofthe prestige he commands in the
community of those competent to judge.
Thus, for advancement within the universities, prestige has a very
high value indeed; it is invoked as evidence for a claim to rewards,
in a situation where it cannot be rigorously tested. Also, the assign-
ment of prestige to a man depends on more than his work within a
field and on the field itself. For universities also have a prestige
ranking; and to be (or even to have been) at a university of high
prestige, automatically confers prestige upon the man. The two
systems of prestige thus interpenetrate and automatically reinforce
each other; and at the top of the scale are those scientists who,
ideally, are at the universities of greatest prestige because of their
enjoyment of the greatest prestige among the expert and lay col-
leagues they have encountered during their career. Once ensconced
in the greatest universities, they will, in principle, continue to work
as scientists of distinction and also function as statesmen of science.
By their contacts with leading scholars of all fields, spiced by the
element of competition which exists even at the rarified heights,
they will master the very subde skills, and develop the very enlight-
ened attitudes, necessary for participation in the government of
science at the very highest level.
This was the ttaditional institutional framework for the ultimate
exercise of quality control and direction in science. Appropriately,
the task of controlling the conttollers reaches its final stage in a very
informal community, operating through a mixture of co-operation
and competition to exercise conttol on each special area in the
interests of a very general common good. The great age of this
system was that ofGermany, in the middle ofthe nineteenth century.
It may be significant that those universities were not merely research
institutions, but were devoted to teaching as well. So while a man
might achieve a fair position by the exclusive pursuit of his research,
he could not join the company of the ttuly great unless he had the
286 Social Aspects ofScientific Activity
breadth of vision and culture to be an influential teacher as well. In
this way, the fragmentation of the community of scholarship was
prevented, or at least delayed. o The functioning of this system was
also improved, and perhaps even made possible, by special features
of its social and cultural setting. The heroic age of German science
came after the period of great philosophical endeavour and educa-
tional reform. To have a genuine university of high prestige was
considered a valuable asset by the citizens of many towns; and there
was available, in secondary schools and universities, teaching which
was both rigorous in intellectual discipline, and philosophical in
approach. For a scientist raised in that environment the highest
achievement was not the making of new technical discoveries, but
participation in ' Wissmschaft': the dedicated quest for human
knowledge, wherever it could be found. 1o
Industrialization ofScience and Quality Control
Some elements of the classic university system (or its analogue in
the centtalized national academies) have been imported into all those
nations where science has flourished since the nineteenth century.
But with the industrialization of science, certain changes have
occurred which weaken the operation of the traditional mechanism
of quality control and direction at the highest level. The concentra-
tion ofthe scientist's audience towards those who advise the agencies
has not merely caused in him a tendency to neglect the diffuse con-
sensus of expert colleagues; it has also made him less dependent on
the judgement of his lay colleagues in the university system. Indeed,
in America, where the process has gone the furthest, the term
'university' has been replaced by a new one, 'multiversity'. In these,
independent research contractors and entrepreneurs grace the
9 On the relation ofthe ideals ofteaching and raearch in ~ German universities, see
H. von Helmholtz, 'On ACidemic Freedom in German Universities', in PoprJ.
Leaw,s tnt S,",,"}i& SIIId,as (London, 11g3), ii, 23~5; especially pp. 252-62. In this
essay, Helmholtz streIIed the conW2'le IIpect of the situation of university scientists:
that university teachers were required to be competent scientists. For the German
system was the exception among universities in the nineteenth century.
10 For a desaipuon and explaDation of the peatDeII of the German universities in
terms of their GU1turaI and social setting, see J. Ben-David and A. Zloczower, 'Univer-
sities and Academic Systems in Modem Societies', E",pe." JOfII'1UIl of SonoloD, 3
(1962),45-84; reprinted in part in N. Kaplan, StietJe, tJtUl Sonety, pp. 62-85. In this as in
other articles, the authors seem to haft a concern to 'reduce' the greablcss to its social
causes, and they tend to explain the decliDe by its inherent features; they do not give
sufficient emphasis to the variable rate of growth through the century, declining to-
wards the end.
 Q.uality Control in Science 287
institution with a share of their prestige, in return for a well-paid
sinecure and the provision of material facilities. I I Under these con-
ditions, the system of assignment of prestige, always a delicate and
unstable thing at the best of times, cannot operate to any worthwhile
effect. The subtle and informal assessments of quality, across fields
and disciplines, and the use of these assessments in distributing
rewards within the university system, are impossible. Once estab-
lished as a leading Big Scientist, a man is immune from control by
anyone except his intimate colleagues and friends, and those in-
vesting agencies whom they joindy advise. He can create property
by the mere flourish of a pen. He may choose to take time off from
organizing to produce some serious research; but otherwise he can
achieve a satisfactory string of pUblication-points (necessary for the
formalities of conttact renewals) by placing hasty, shoddy papers in
journals controlled by himself and his friends. 12 If such men come
to dominate the government of a field, the effects, especially in
times of financial stringency, are easily seen to be disastrous.
The extent to which such developments have already damaged
the mechanisms ofquality control in science is impossible to estimate.
The evidence on which such an estimate would need to be based
would have to be worked up carefully from very private and con-
fidential testimony. For our present purposes, it is sufficient to
appreciate that some such things have occurred, and that they are the
effect of changed technical and social conditions on a very delicate
task. In the last resort, there are no direct means whereby anyone
outside the world of science can exercise quality control on science.
The products of the craft work of scientists are intelligible, and
valuable, only to other scientists. And although they relate to the
external world, their value as well as their meaning is governed by
the judgements of men, those particular men who enjoy this esoteric
activity. If the government of this work were accomplished through
II A disturbing description of this situation, orpnized around the problem of student
unrest nther than of the gov~t of science, has been given by Paul Doty, 'The
Academic Condition in the United States', NIIIW" 224 (1g6g), 1063-8. The author is a
Professor of 01emistry at Harvard; what he says of that aristocratic institution can be
expected to hold more strongly elsewhere.
12 Evidence that the accumulation of the pseudo-property of publication points is not
restricted to those that can do DO better, comes· from a letter by R. W. Whalen in
Seime" 16, (1970), 13 18. He reports an author who first published a paper in Seime,
(itself a mark of quality), but who proceeded to use one-half of the ckta, and then one-
quarter of the data, for two later publicatiODl in speciality journals, with DO aoa-
referencing between articles.
288 Social Aspects ofScientific Activity
formal institutions, then its response to changing conditions would
be delayed, and the work might have time to adapt itself gradually
while maintaining its excellence. But the nature of the work requires
a government for direction and quality control which is almost
entirely informal, accomplished by a series of craft skills which
become ever more refined, demanding and delicate; and the work
itself is very sensitive to the quality of its government. The problem
of quality control in science is thus at the centre of the social
problems of the industrialized science of the present period. 13 If it
fails to resolve this problem, and does not develop new techniques
for restricting prestige and rewards to those who deserve them, then
the immediate consequences for morale and recruitment will be
serious; and those for the survival of science itself, grave.
II The problem of quality became a practical CODCCI'D for American scientists when
the JoImIoa administration supported. the policy of establishing new 'centra of excel-
lence' in science away from the areas which bad come to monopolize prestip and funds
(the North-East mel California). It was appropriate for • representative of the molt
aristocratic dilcipliDe, mathematics, to look at this problem of '1Cience policy' from the
point of view of the maintenance and control of quality; he showed that 'excellence'
cumot be apandecl merely by a shower of dollars. Sec S. MadaDe, 'Leadership and
Qpality in ScieDce', in Billie R,• •ell all NIIIiouI GtMh (A Report to the Committee on
Science and Astronautics, U.S. House of Representatives, by the National Academy of
Sciences) (WuhiJIston, Government, Printing Office, 1965). The paper was reprinted in
put u 'Can we Buy QpaIity in ScieDce 1', BII1Jeti" '11M AI.ae S&inllists, :&1 (November
1965),7-11. Britishscientistsbave a1Io become mncemed about the problem of quality.
In addition to the 'lCendrew Report' for molec:ular biology, there was an article by the
_ physicist A. B. Pippud, in which quality control bad a brief but sipifi-
cant mention. In 'ScieDce u • Constituent of University Education', NMw', 219
(1969), 1307-'8, Pippud made various proposals for rationalizing research and teaching
in physics; but confcaed that 'the problems involftd in encouraging centra of excel-
lence and, perhaps more important, discoungins mediocre research are too complex to
diIcua here.,
 II

ETHICS IN SCIENTIFIC ACTIVITY

ON this very important topic, we need to define our problem very

carefully; otherwise we are sure to encounter some of the many pit-
falls that beset any discussion of the relations between what is and
what ought to be. We can start with the phenomenon of the sudden
shift in the public image of the morality of science. Through the
academic period, science was portrayed, and widely (although not
universally) accepted as. an activity good and ennobling in itself, and
productive of enormous although diffuse benefits for mankind in
general. In recent years, the image has been tarnished; not only is
~cience blamed for the new horrors of war and the threat to our
environment, but there is a new genre of literature exposing the
human foibles ofscientists. I It is all too easy to conclude that science
as a whole has declined from a state of pristine purity; but such
popular and popularizing materials are extremely unreliable evidence
for the state of affairs which they are taken as describing. If nothing
else, they can deal with only a very small segment of a large and
variegated activity; and whether this is truly representative can be
determined only by an extensive study in depth. This too has its own
hazards, for any relevant historical data is inherendy unreliable; and
without an appreciation of the special character of the problem, the
most diligent research could easily yield vacuous results. Thus, a poll
of elderly scientists would be quite likely to produce a majority view
that in the old days scientists were more honest and self-sacrificing,
and also that summers had more sun and winters more snow.
I The frank display of human motives in James D. Watson, Tile DouJle B,u: II
Persoul AttOflftt of the Distover1 of DNA (Atheneum, New York; Weidenfeld and
Nicolson, London, 1968), was a source of revelation for many reviewers. A very different
reception was given to D. S. Greenberg, The Politits ofP", Slime,: the leaden of the
American scientific community were DOt amused by the story of the human reactions of
scientists when faced with the temptations of money and power; nor was the book's
message clearly perceived by reviewers in the general press.
290 Social Aspects ofScientific Activity
Our present task, then, is not to decide whether the two dominant
images of science are correct portrayals of the past and the present;
but to establish the categories in whose terms problems relating to
that question can be properly investigated. Although such a study can
be only a preliminary to a substantive enquiry, it is not merely an
abstract exercise, and some general conclusions about the real world
of scientific activity can be derived from it. This present discussion
will be organized in terms of two problems. First, in what ways, if
any, does the special chara\,1:er of scientific inquiry necessitate the
adoption of a superior ethical code by scientists, for good work
to be done? Second, to the extent that this is the case, then by
what social mechanisms and individual actions and commitments is
this ethical code enforced and maintained? In both of these ques-
tions, we will use materials developed in the preceding chapters,
and amplify certain points mentioned in passing there. Also, we will
assume that it is possible for good work to be distinguished from
bad. This assumption is justified by the experience ofmatured fields
ofinquiry; in this discussion we can ignore the problems ofthe iden-
tification of good work, and restrict ourselves to using the general
principle that doing good work is more costly to the individual
than doing bad work.

The 'Pinal Causes' ofa Task

We have already mentioned that one function of the various
social tasks in science is to maintain a harmony between the collec-
tive goals of the endeavour, and the private purposes of individuals.
For our present discussion, this distinction between the 'objective'
and 'subjective' final causes ofan activity must be enriched and made
more precise. For this, we will develop a set of terms, using words
taken over from common speech and nearly synonymous there, but
which we will henceforth use in the restricted senses to be defined.
Our unit of analysis is a task on which someone is engaged; and for
the sake of clarity we will take as our example a task where the
various causes are very distinct. z
Suppose we find someone on a routine, assembly-line job, say
fitting together a set of components into a sub-assembly. If we wish
a The literature of the sociology of orpnizaticms distinguishes among the various
'golla' involftd in such work, but the present classification is not to be found there. For
a recent study see E. em-, "The Definition of Organisational Goals', Brit;sh JounJIIl
,/SotioltJo, 20 (1g6g), 277-94-
 Ethics in Scientific Activity 29 1
to know the 'aim' of his task, its final cause, we find it necessary to
put four quite different questions: 'What use will it be ?', 'What is to
be done here?', 'What are you trying to do?', and 'Why are you
doing it ?'3 The first question relates to the function that will be
performed by the product of the accomplished task. In a system of
production with a division of labour and a hierarchical organization,
this part of the final cause of the task is the most basic. For, from
the pre-assigned function, the goals of the task are specified; the
fulfilment of the goals ('What is to be done ?') is the means to the
performance of the function. These goals include the specification of
the accomplished task, and criteria for its adequacy. In the assembly-
line example, these will relate to the properties of the product which
determine its quality (tolerances of fit, shape, etc.), and those of its
making which relate to its cost (as materials, wastage, and time
required). Now, the function which determines the goals of the
task does not exist in a vacuum; it will generally be subsidiary to
other functions in a hierarchy. The ultimate 'final cause' at the top
is not a function, but a set of personal purposes, individual or col-
lective, directly expressed or imputed; for otherwise no one would
have brought the activity into being or kept it going.
We notice that these first two questions are not best answered by
the operative himself, but by those superiors who define, direct and
control his actions on the job. Only in the last two questions do we
come to a man, as distinct from a biochemically activated machine.
A machine may be designed to direct its motions and processes for
the fulfilment of preassigned goals, but the man has purposes. These
are what he wants to accomplish in a particular task; those things
that once done, will give him satisfaction and the sense of effort well
spent. The fulfilment of the assigned goals of the task may well be
among his purposes, either for their own sake, or more probably
as a necessary condition for the achievement of a genuine purpose as
entitlement to the day's wage without penalties or deductions. This
distinction between the operative's purposes on the one hand, and
the goals and higher final causes of the task set him on the other, is
fundamental to this argument. There are still many who believe that
the subordinates in a hierarchical organization (be they workers,
soldiers or students) either do, or should, or in the good old days did
3 For a closer study of some of these 'teleological' concepts, applied to sociological
theory, see Dorothy Emmett, Flm&htm, Pwpose flU P01lJtrs (MacMillan~ London; St.
Martin's Press, New York, 1958).
292 Social Aspects ofScimtijiG Activity
have their own purposes defined and dominated by the final causes
imposed on their tasks: that they do, should, or did believe in their
boss and sincerely want him to be pleased with them. Without
arguing this set of views, we can agree that, as a contingent fact,
many workers, soldiers and students now do not particularly
believe in their boss, and want only to stay out of his way: their
private purposes are not dominated by the imposed goals of their
tasks. Whether such a situation is new or old, good or bad, is not of
concern at the moment; it is necessary only to establish that there is
in principle a difference between the objective final causes imposed
on a task, and the peisonal, private ones of the operative. Once we
accept this difference as existing somewhere, sometime, then we
have to reckon with the problems ofquality-control and ethics in any
socially organized activity including scientific enquiry.
Coming to the last of the questions defining the final causes in-
volved in the task, we consider why the operative is at that particular
task (to the extent that his situation is the result of choice) rather
than at some other. We call this set of reasons his 'motives'; and we
notice that these set part of the framework in which his private
purposes are specified. The motives will include some desirable and
some undesirable aspects of the situation; and the purposes will
correspondingly include the maximizing of certain personal benefits
and the minimizing of certain personal costs, either specific to that
situation or of a general nature. To take a trivial example, one may
find two men next to each other on an assembly line; one may be
there because ofthe opportunities for overtime on that job, while the
other lives close to the factory and works there so as to have more
time for his leisure activities. The purposes of the first will then
include the achievement ofa significant amount ofovertime pay, and
he will do what he can to influence the flow of work so as to make
extra hours necessary; while the other will have a contrary purpose,
and his private strategy on the job will include seeing that the week's
work is done in good time. 4
The specification and achievement ofa set ofpersonal purposes on
a job is thus a complex affair, which can itself be considered as a
task. The different sorts of activities and unit tasks associated with
.. My ~ of the four tams 'function', 'goal', 'purpose', 'motive' is to some extent
arbitrary, and even u I have defined then hae they will overlap in meaniDg with other
COIDIDOIl tams clcDotiDI 'coda'. I have .voided 'objective', since it generally meaDS
'goal' with a connotation of 'function',and 'aim' or 'aims' which mix 'goal' with 'purpose'.
 Ethics in Scientific Activity 293
a job are evaluated by their function in the achievement of the per-
sonal purposes, and individual unit tasks will be specified by private
goals, established in accordance with these private functions, rather
than being defined by those who direct and conttol the work. Thus
we may see someone working quite hard at the accomplishment of
some assigned task, but what he is 'trying to do' is defined by his own
private system of basic purposes, subsidiary functions, and in-
dividual goals. The goals set by the organization operate only as
external constraints on his specification of the complex private task;
and these constraints are effective only to the extent that. his fulfil-
ment ofthe goals can be conttolled. Since quantity is easily measured
and (in this situation) quality is capable of assessment by routine
tests, the two systems of final causes can coexist, if not in hannony
then at least in relative peace.
The difference between a man and a biochemically activated
machine introduces another element in the government of his
activities: one which is social and in part ethical. Since the operatives
necessarily have activities other than their assigned manipulations,
the organization must impose a job-discipline, governing such things
as punctuality, rest-breaks, and general behaviour on and off the
production line. Without this, the shop would degenerate into chaos
and the imposition of conttols, even on the work being done, would
be severely hampered or impossible. But those on the shop floor
have a society of their own, with its own code of behaviour. This has
the function of protecting the common purposes of its members,
and will frequently run counter both to the system ofconttols and to
the job-discipline. It will frequently require a member to forgo
personal advantage for the sake ofthe common good, as in refraining
from exceeding the standard rate ofproduction, or in joining a strike.
In order that members should do this voluntarily, they must sub-
scribe to an ideology from which this ethical code can be derived,
and which serves to justify and guide their actions in this respect.
In its terms, struggles with the organization are conceived as the
protection of their rights.
In this extreme case, the radical separation of the organizational
and personal components of the final causes of the task creates a
situation where the worker's ethical principles always operate against
the fulfilment of the organization's goals. Naive appeals to the
'common interests' of all ceo-workers' in the firm (be it privately
owned or part of a Socialist economy) will inevitably be greeted with
294 Social Aspects ofScientific Activity
public or private derision. 5 From management's point of view, the
operative cannot be expected to show initiative or be entrusted with
responsibility; his advantage over a machine in the subtlety of his
manipulations is offset by his unreliability on the job and his cunning
applied against the job. For the worker, convenience and right
combine to encourage him to get as much as he can for as little, by
any practical means at his disposal. Of course, there are many cases
in which enlightened or skilful management succeeds in overcoming
the harshness of this dichotomy; but the raw situation is sufficiendy
common to serve as an example of a real case. And taking this
extreme case as a real and in many ways natural situation, we can ask
how it is that scientists can generally strive for the achievement of
the results of the highest quality of which they are capable, in the
absence of a formal system of quality control or of any imposed job
discipline. Either they are drawn from an elite group of recruits of
great moral superiority; or some aspects of their task, and its in-
herited code of behaviour, enables them to develop very enlightened
private purposes which harmonize with the common good.
!l.ua/ity Control ofSleilled Tasis: the Insufficiency ofSelf-Interest
We have seen that in the case of the unskilled production-line
worker, his private purposes and his ethical code are basically in
5 This initial naiv6t6 about the behaviour of the working class under socialism may
we1l have been responsible for lOme of the worst excesses of Stalinism. A theme which
IUDS through A. Koestler, D"'ftess III Notm, is the contradiction between the aspintions
of Socialism and the intellectual and political backwardness of the Russian masses;
indeed, this provided the rationale for the 'confession' by the bourgeois-intellectual
Rubashov. The peasant interrogator Gletkin provides the evidence for the problem,
first in teI1ing of his discovery of the necessity for coercive methods in Socialist con-
struction; and at the end, with the question 'Were you given a watch as a boy?' (penguin
Boob, 1947), pp. 80-8 and p. ISo. At least some of the more recent revolutionaries have
publicly appreciated this problem; thus Che Guevara discussed the difficulties caused
by the ideological backwardness ofthe Cuban urban working class. See 'On Sacrifice and
Dedication' (1960) in VmetrtllUJs, cd. J. Gerassi (MacMillan, New York, 1968). Lenin's
own teaebiDg on the problem was not very helpful. In StilI' IIU ReTJOlut;o1l he gave the
classic formula: that the repressive state would gradually wither away spontaneouslyt
when 'freed from capitalist slavery, from the untold horrors, savagery, absurdities and
infamies of capitalist exploitation, people will gradually become fK&UStometl to observing
the elementary rules of social intercourse' (ch. 5, section 2, p. 462). But shortly after-
wards he recognizes the necessity for maintaining 'the stri&I,SI control by soriety lind by
tile stilt" (italics original) so long as the 'ordinary run of people, who like seminary
students in Pomyalovsky's stories, are capable of damaging the stocks of public wealth
"just for fun" and ofdemanding the impossible'. (Ibid., ch. 5, section 4, pp. 469-70; the
order of the passages quoted here is inverted from the originaL) Page references are to
the Colletted Woris, voL xxv (Lawrence and Wishart, London, 1964).
 Ethics in Scientific Activity 295
antagonism to the fulfilment of the goals set by the organization
which employs him; and he fulfils these goals adequately only
because his work and activities can be controlled and disciplined. If
we now consider skilled work, we find a radically different situation.
For the quality of work cannot be controlled by routine tests; and if
good work is to be done, it is necessary for the operatives to want to
do it. And although, as we will show, maintaining the quality ofwork
performs a valuable function for any group of skilled workers, this
in itselfis not a sufficient reason for anyone ofthem to exert himself
to that end. Hence something more is necessary, as a component of
the personal purposes of the craftsman on his job.
I mentioned earlier, in connection with quality control on skilled
production work, that the assessment of the quality ofsuch products
cannot be done on a routine basis. This is only a special case of a
more general principle, which has direct relevance to scientific work.
In the measure that goals of a task are complex, sophisticated, or
subde, then the assessment ofthe adequacy ofthe fulfilment of those
goals must be correspondingly so. For if the tests of adequacy are
crude in comparison, then a person possessing the skills to do the
work properly will be able to find ways of formally satisfying the
imposed tests while achieving his own purposes to the detriment of
the assigned goals. This holds for the production and management of
material things; and when the product of a task is immaterial, as in
organizational or managerial tasks, evasion of the intended function
of any system of conttols is child's play. This phenomenon has been
recognized in nations with centrally planned economies, where
individual factory managers make certain to fulfil the bureaucrats'
goals for their establishments, and yet the 'plan' never achieves the
desired co-ordination of its parts. And from the advanced society at
the other extreme of social organization, America, comes the
aphorism, 'Wherever there's a system, there's a racket to beat it.'6
This problem is not resolved by imposing more complex sets of
6 For this aphorism I am indebted to an IDODymoUS tram-driver in Atlantic City, New
Jersey, in 1952 • There was system which made it impossible for him to cheat on the
collection of fares: passengers put their fares into • machine, and he bandled money
only to make change when necessary. But the fare was nine cents or a token worth ODe-
third of twenty-five cents. On a Saturday evening he had many couples boarding the
tram for this occasion only; the man would offer a twenty-five cent piece. The driver
would give change of seven cents, as from two full fares, and drop two tokens into the
machine. The profit was only one and a third cents each time, but several hundred such
fares amounted to a tidy bonus for an evening's work. The driver was euphoric over his
discovery and shared the secret with nearby passengers, along with the valuable aphorism.
296 Socilll Aspects ofScientific Activity
controls on the work. For the system ofcontrol may itself become so
cumbersome as to invite close analysis and really determined
attempts at complete evasion; and the problem of controlling the
conttollers can also become acute.
It seems that this problem, which is closely related to the general
social problem of corruption, can be resolved only if the agents
being controlled acquiesce in the operation of the system and also
participate in it. How this can be done with a general public is a
most difficult problem; but we have many examples ofspecial groups
which do in fact police themselves to a 1arge extent. In the case ofan
occupational group, such as skilled workers or professionals, the
mechanism seems straightforward. All the members of the group
enjoy special rewards in virtue of their membership; but these exist
and are maintained only because membership in the group is a
guarantee, to employers or clients, of the high quality of the work
which will be performed by anyone so designated. If this guarantee
is betrayed, the reputation of the group will be lost, and it will lose
its special rewards, either gradually or with catastrophic suddenness.
Thus in this case, unlike that of the unskilled worker, each task
accomplished serves as a precedent for the support of the future
claims of the individual worker, as a member of his group, to special
rewards. Hence, we might reason, each member will naturally strive
to maintain that common reputation, out of simple self-interest.
Arguments of this character, showing how a common good can
derive from an aggregate ofselfish decisions, are as old as the science
of political economy. They all share the same logical fallacy, and all
neglect the incommensurability of social cost and private benefit
which can frequendy occur in human actions. 7 The fallacy lies in
assuming that 'all' and 'each' are completely synonymous and
interchangeable logical operators. An argument, like the above,
which starts with 'all' and ends with 'each', is not necessarily valid.
To show this, I will use an example involving the incommensura-
bility principle, and suppose that I discover a way of stopping my
electricity meter. If my manipulation does not leave any traces, I cut
it off between visits by the meter-reader, and I reduce my bill by an
, An IIpIIDalt similar to mine has been made by M. Olson, in The Loti' DfColJeaiw
AnUm (Hanard University Press, 1965). He shows there that individual actions which
are 'ratioDal' in the economists' seDIe of the term will, in poeral, be mUDtel' to the
acbicvement of their share of the common purposes of their poop. He is mainly mn-
ccmed with the implicatioDs of his coadusioDs for public policy; the ethical aspects of the
problem are eliminated by the concept of 'rational'.
 Et"ics in Scientific Activity 297
amount which would not attract suspicion, why should I abstain?
I can argue that even if my gain were recouped by the company
through slighdy increased charges to the other consumers, the loss to
anyone of them would be quite negligible; and so the total sum of
'happiness' or 'utility', measured by cash in hand, would experience
a net increase I In opposition, it could be argued that ifall consumers
tapped the meter in this fashion, then the rates would rise until the
effective tariff to each would be the same as previously, and the
community would be the worse for the universal cheating. But this
reasoning fails to be binding on each of the consumers, and in
particular on this one. In answer, I could take two cases: where the
dishonest consumers are a negligible fraction of the total, and where
they are a significant fraction. The former case is the same as that
where I am the only beneficiary; and in the latter case, I could argue
that I would be a fool to pay significandy more than my due, to
reward those other consumers for their dishonesty. Nor is the situa-
tion changed if there is a risk of detection; I could simply work out a
probabilistic cost-benefit function involving the penalty, proceed
somewhat more circumspectly, and argue the principle ofthe matter
as before. And although it is clear that if each member of the group
yields to temptation in this way, all will soon suffer, still there is no
argument from utility to all, to prove to each man that he should take
the straight and narrow path.
It seems, then, that situations where private benefit from an action
is incommensurably larger than its social cost, are not conttolled by
the application ofa 'utilitarian' ethic, nor by the principle ofaggrega-
tion of enlightened self-interest. The existence of this class of
apparendy harmless aets of dishonesty is recognized in the English
vernacular, in the term 'fiddle', as distinct from theft or embezzle-
ment. What ethical grounds, if any, are there to deter me from
tapping my electricity meter? There seem to be two. The first is the
very absttaet form of utilitarian doctrine, deriving from Kant: that I
should judge my actions by the consequences that would result if
everyone else were doing the same. And the second is that certain
things are simply not done, regardless ofwhether anyone else knows
of them or imitates them" and independently of their actual or
hypothetical consequences.
This ethical problem, and its possible solutions, are directly
relevant to the maintenance of good work in skilled and sophisti-
cated tasks, including science. If I am the only effective judge ofhow
 298 Social Aspects ofScientific Activity
well I am fulfilling the goals of my task, why should I refrain from
cutting comers here and there, and helping myself to a bit of
personal benefit within my reach, when I am quite sure that no one
would notice or even particularly care? On what basis should I
decide to make the personal sacrifice involved in doing my work
better than I need to? Any imposed criteria of adequacy on my task
and job discipline, are quite irrelevant to my decision on this matter;
nor am I constrained by a group code of behaviour designed to
maintain the reputation of the work of all. For me to include among
my personal and private pwposes, the best possible accomplishment
of my assigned tasks, I must be conditioned by two closely related
principles: loyalty to my group, so that I would not let the side
down; and pride in my own work, so that I would not betray
myself.
These two principles can be effective on individuals, only when
there is good morale in the group; indeed, the presence of such
principles is an essential component of good morale. In its absence,
work of excellence will not be done by all; and it is hard for any
one individual to maintain such work for any time, in isolation from
and in opposition to his colleagues. The achievement of good morale
is not a simple task; imposed systems ofrewards can help, but if they
do not perform their intended function, they merely become the
occasion for a new system of cheating. It is an error to believe that
the commitment to good work, and the associated good morale, can
arise automatically because the social system of the accomplishment
of the tasks will thereby provide an absolute maximum of benefit to
all. This is the fallacy of 'functionalist' explanations of human
behaviour, which (unlike Aristode) assume that the 'best' final cause
is simply identical to the efficient cause. It is exceedingly common
for such social systems to achieve a relative maximum of general
benefit by an accommodation rather than a harmony of collective
goals and private pwposes: where the private purposes of all the
agents (including those exercising decision and control) are domin-
ant over the collective goals, and where these latter are fulfilled in a
perfunctory or spurious fashion. To bring group activity from such a
slack and demoralized state, to one of excellence and commitment,
requires leadership capable of establishing an ethic, based on an
ideology, to which the members adhere. For the ttansition is a
highly unnatural one, and even the maintenance of the state of
excellence requires the continuous exercise of leadership.
 Ethics ;" Scientific Activity 299
The Management ofProperty: the 'Professional EtitJuette'
It might appear that the ethical problems identified in the case of
skilled work are not directl}~ ~elevant to the case ofscientific inquiry.
For the tasks of the skilled worker differ from those of the unskilled
only in the degree ofdifficulty oftheir control; the ultimate purposes
from which they are derived are still irrelevant or antagonistic to his
own. Hence it is only natural for the normal state to be one of con-
flict between organizational goals and private purposes, and for any
true harmony to be the exception. In science, by contrast, there is
no such basic antagonism; for there the situation is one ofa group of
colleagues freely co-operating in a common endeavour, and where
the social tasks of control and even of direction are undertaken
informally by a group within the community, differentiated only by
their talent and commitment. A mark of the fundamental difference
between the two cases is the absence of an imposed job discipline,
and the refinement of the code of behaviour into an elaborate
etiquette whereby the distribution of rewards for achievement is
accomplished with honesty and fairness.
Yet these particular differences do not necessarily result from a
uniqueness of scientific endeavour, whereby it is essentially and
automatically governed by the highest ethical principles. As we have
already seen in our discussions both of the protection of property
and of the management of novelty, scientists have private purposes
which are not always necessarily in harmony with the collective
goals of the work. It is true that autonomous research scientists are
not in the work-situation of an employee. The product of the task is
not a precedent for the assessment of quality made by an
external agent of control; it belongs to the scientist himself,
as a piece of intellectual property. But there are other spheres
of activity where a similar sort of property is created, and where
an etiquette can function quite independently of refined ethical
principles.
One essential feature of the scientist's intellectual property, as
distinguished from the 'real' property of commerce, is that it exists
only by being available for use by others. Even if the scientist does
his research in physical isolation from colleagues, his tasks are
eminently social in nature. He depends on others, not merely for the
training in his craft skills, but also for the information without which
he could not do any work at all; and his property takes on value only
300 Social Aspects ofScientific Activity
through the explicit assessments of colleagues, rather than through
the workings ofan anonymous market. Now, there are other sorts of
work, which have similar features of essential co-operation and
mutual assessment between independent practitioners, although dis-
tinguished from science in that the individual task usually brings a
cash return on its own. The ttaditional professions are examples of
this class of work, and in them the 'professional etiquette' performs
a very straightforward function in the harmonizing of private pur-
poses and collective goals. 8
In professionals' work, the assignment of shares of personal
property is done more in the definition and division of the task, than
in the acknowledgement of previous work used in the current pro-
ject. But in spite of this difference from science, the function of the
etiquette, and the reasons for its necessity, are the same. In the
learned professions in particular, the task presented by a particular
'case' may require a variety of skills for its accomplishment. The
practitioner who receives such a case must decide whether to try to
do the job on his own, or whether to call in colleagues expert in
particular aspects, sharing the work and the payment with them. It
is clearly in the clients' interest to have competent handling in all
respects; and the collective purposes of the profession are served by
ensuring that this will always occur. Hence the task will be parcelled
out among various experts, each sharing in responsibility for its
accomplishment, and enjoying a part ofthe reward. But this division
of the task is a highly sophisticated expertise; the analysis of the task
itself constitutes a problem like a scientific one and then the choice
of fields of expertise and the division of responsibility among them
requires a skilled knowledge of both the technical and social aspects
of the task. The control of this division of the professional's task,
presents the same difficulties as the control of quality of the work:
no formal set of rules could encompass the variety of situations en-
countered, and a machinery of decision or adjudication would be
impossibly cumbersome. Hence for this division to be accomplished
quickly and efliciendy, it is necessary for the practitioners to adhere
to an informal code of behaviour, the mastery of which is one of the
more subde skills necessary for competent practice in the field. This,
then, is the essence of the professional etiquette: the methods of
• For a recent summary of the literature on 'professionalization', including clusic
lOUI'Ce8and modem research, see P,oJiuiDuli~ cds. H. M. Vollmer and D. L. Mills
(Prenticc-HaI1, 1966).
 Ethics in Scientific Activity 301

social behaviour necessary for the orderly division of the property

created in the particularly sophisticated co-operative work by inde-
pendent agents.
Although a professional etiquette will necessarily be more
elaborate and intellectually subde than the code of behaviour of
skilled manual craftsmen, it is not thereby endowed with a higher
ethical content. Indeed, an aggregation ofself-interested policies can
suffice to maintain such an etiquette. For ifone member ofthe group
violates the etiquette by failing to call in a colleague from the appro-
priate group of experts, he is not merely endangering the reputation
ofall the members ofthe profession; he is also appropriating some of
the payment which that group would expect to flow to a member.
By a combination of collective goals and private purposes, they will
be motivated to impose penalties on him; and so in this respect the
system provides an automatic mechanism for enforcement. Also, it is
possible to identify codes ofbehaviour which function as professional
etiquettes in situations where refined ethics are definitely absent. As
an extreme example, the code of the American Mafia is said to con-
sist of two maxims: don't squeal to the cops, and keep your hands
off other members' wives. Although the truly learned professions
would not have such crude etiquettes as these, there too they can
operate exclusively for the division of property, to the detriment of
the interests ofclients. American medicine occasionally has scandals
from revelations of unnecessary consultations with kick-backs of
fees; and in any profession the etiquette operates to close ranks
against aggrieved clients. Even in science itself, one can find stagnant
fields, where the creation and division of intellectual property is
governed by the most rigorous etiquette except in those cases where
genuine work challenges the existing social structure of the
field.
By itself, then, a professional etiquette does not establish a har-
mony between the private purposes ofpractitioners and the professed
ultimate purposes of the work. If the collective goals of the activity
are conttary to those ultimate purposes, that is, serving an aggregate
of interests of members instead of those of clients, then the
etiquette can operate quite directly against the achievement of
those ultimate purposes. ,For these ultimate purposes to become
relevant to the definition of the final causes of the professional's
task, the etiquette must be supplemented by a genuine professional
ethic.
302 Social Aspects ofScientific Activity
Professional Ethics and their Enforcement
A true learned profession can be distinguished from an occupa-
tional group which provides a particular service by four criteria.
The client is dependent for his welfare on the accomplishment ofthe
task; but he is not competent to assess the adequacy of the work
done; recognized competence in the set of tasks is legally restricted
to those certified to have completed a training of a scientific char-
acter; and in exchange for the monopoly of practice the group
accepts responsibility for the achievement of the purposes of clients.
The situation of the professional thus involves an essential fiduciary
element; incompetence or malfeasance constitutes a bettayal of the
client's trust. Should this occur, there is a risk of scandal, and the
erosion or loss of the legally enforced monopoly enjoyed by the
professional groUp.9 It is thus in the long term collective interest of
the profession to maintain standards of work and to protect the
interests of clients; but this does not automatically extend to the
immediate private purposes of the members of the group, either by
self-interest or by the operation of a professional etiquette for the
orderly division of property. A more refined code of behaviour is
necessary to prevent the relaxing of standards and the neglect of the
interests of clients, whereby the profession would degenerate into
a group of businessmen temporarily enjoying a collective monopoly
on a particular class of ttade.
This social function of maintaining the health of a profession is
performed by a professional ethic, as distinct from an etiquette. IO
This ethic must be based on the absolute primacy of the true
ultimate purpose of the task: serving the purposes of the clients.
This need not mean that the private and collective purposes of the
profession should invariably be sacrificed in the interests of each
• In the earlier nineteenth century American orthodox mediciDe suJrered from such
ampetition &om sedUiaDs and quacks, partly because the professioDal physicians were
pftll to 'heroic' treatments (the heroism being that of the patient enduring copious
blood-lettins or poisonous purgatives). Their reputation sank so low that some states
revoked their monopolies. Most of the sects combined a 'democratic' style with mild
remedies, thereby achieving lasting success; but the COIDIDa'daI 'patent medicines' were
sometimes even more deadly than the physicians' prescribed remedies. See R. H. Shryock
T1Je DewID".", ofMotlmJ M,tlici. (Knop~ New York, 1936; Gollancz, London, 1948),
m. XIII, 'Public Confidence Lost', pp. 22.i-48.
10 The distinction between professioual etiquette and professional ethics is clear in
the pandipn code of medical ethics published by Thomas Percival of Manchester in
1803. The latter was 'reviewed' in the code, while the former was given detailed treat-
ment. See R. H. Shryock, T1Je DewID".", of MotlmJ M,tlid_, pp. 138-29.
 Ethics in Scientific Activity 303
individual client; the protection of the interests of clients as a class
may well involve giving less than full satisfaction to some particular
case. But as a general guiding principle, to be interpreted with skill
and subtlety in each particular instance, the professional ethic is
necessary if the fiduciary relationship is to be justified and pre-
served. The problems of establishing and maintaining such an ethic
are similar to those encountered in quality control of skilled work,
but even more delicate and demanding. For the control is not merely
on the tasks accomplished, but also on the subtle social mechanisms
whereby the tasks are organized and rewards distributed. The pro-
fessional etiquette itselfmust be governed in its operation, and con-
standy modified as circumstances change. The assessment of
adequacy can relate as much to the way things are done, as to the
fulfilment of stated goals; and the relevant criteria can be so in-
formal as to be incapable of verbal argument. The area of ambiguity
between proper and improper conduct is broad and ill-defined, so
that an individual can claim (perhaps with sincerity) that his parti-
cular bit of self-serving is neutral or even beneficial with respect
to the collective goals. And when all the prior phases of the task of
control are necessarily so extremely informal, a system of crude,
quantitative penalties would find no middle ground between those
that are ineffectually light, and those that are harshly severe.
For the effective enforcement of a professional ethic, it is not
necessary that all members of the profession adhere to it with a
total commitment. But even more than in the case of skilled tasks,
there must be excellent leadership and good morale. The leadership
is necessary to provide definition and application of the professional
ethic in constantly changing circumstances; and morale is necessary
if this leadership is to be effective. For the ordinary members of the
profession must respect the leaders in their own terms of skill and
success, and be prepared to accept their guidance and control. Also,
they must have some degree of commitment to doing good work, so
that a consensus can operate against those who betray the good name
of the profession. Otherwise the enforcement of a professional ethic
is impossible. Even the formal institutions of statute and judicial
law cannot operate effectively (except in conditions of police terror)
in areas where the population at large does not accept its categories;
and in this very informal situation, where the policing must be done
by the members themselves, general acquiescence in the fairness and
necessity of the penalties is absolutely essential. Also, since the
304 Social Aspects ofScientific Activity
whole operation of conttol, including penalties, is generally con-
ducted completely informally except in the most extreme cases, the
subde penalties of disgrace and even disapproval must be real for
the ordinary members ofthe group, if control is to be at all effective.
And for this, there must be good morale: a consensus, applied to
oneself as to others, that good work and a good reputation are
important.
Thus the professional ethic can operate only within certain con-
straints. If the ideology on which it is based (a conception of the
profession and its service to its clientele and the community) loses
contact with the reality experienced by members of the group, the
basis ofrespect for the ethic will erode, the leadership which adheres
to the ideology will be isolated, and the task of policing will become
onerous and eventually impossible. Hence the professional ethic
and its ideology cannot rise too high above the working etiquette,
and the experience and values it reflects. On the other hand, the
situation of the professional makes it natural, in some respects, for
the professional ethic to be comprehended and accepted. The basic
task is related to particular clients whose needs are visible and real;
and the trust of those clients, and the advantages of a protected
monopoly of practice, is part of the daily experience of the practi-
tioner.

lewis ofEthical Commitment in Science

We are now in a position to analyse the problems of the nature of
the ethics and ideology necessary for the maintenance of a healthy
state in science, and of the conditions under which that ethics and
ideology can be effective. It would be unrealistic to assume that a
man automatically becomes a paragon of virtue on receiving his
Ph.D.; and yet the delicate and subde character of scientific work
makes it particularly susceptible to the degeneration of quality in the
absence of controls. We have already seen that in spite of the in-
formal structure of the scientific community, there are several quite
definite social functions to be performed, with some differentiation
within the community on the basis of the accomplishment of the
relevant individual tasks. We can use this rough division to explore
the degree ofethical commitment which is necessary for their proper
accomplishment, keeping in mind that each scientist is not included
in one single group associated with a particular task.
At the lowest level ofthe structure ofcontrol, we have the investi-
 Ethics;" Scientific Activity 305
gation of a particular scientific problem. Considering this as a task,
we can say that its goal is the achievement ofa result. The function of
the accomplished task is to provide materials for the further investi-
gation of other problems; this function may have any degree of
specification from the closest to the most general, depending on the
context of the investigation. Since scientific inquiry is very skilled
and sophisticated work, a scientist who will do only what he is told,
and that not very well, will produce nothing much better than
vacuity. Even a technician assisting on scientific work must have
some personal commitment to the quality of his products if they are
not to be utterly unreliable, even if he is closely supervised by his
superior. Of course, the social context of scientific inquiry is such
as to tend to eliminate in advance those who make the most simple
and short-term calculation of personal cost and benefit; the lengthy
process of obtaining the necessary skills and qualifications requires
an attitude which, if not necessarily more enlightened, does require
a more sophisticated and long term calculation. However, this
process does not exclude a man who supplements his research skills
with those of getting by with least work. In principle, the quality of
his work is controlled by the referees of the papers he submits to
journals; and they should have the special skills for detection of
slipshod work. Thus, for the maintenance of a minimum standard of
competence in the productions of ordinary scientific workers, it is
not necessary for them as a body to have a degree of commitment to
good work any greater than that ofskilled manual workers, provided
that they can be effectively policed by referees.
Higher demands are made on the referees, if they are to do their
work properly. If, out of indolence or personal interests, they all
pass everything that looks plausible, then the public channel of
scientific communication in their field would quickly become choked
with an undifferentiated mass of results ofevery degree of quality,
mostly bad. For their work, which is frequendy onerous and which
brings no tangible rewards, the referees must be committed at least
to the maintenance of the standard of quality in that field with
which they are associated. But for this degree of discipline, it is
sufficient in the short run that the referees adhere to a code of
behaviour equivalent to a professional etiquette. The value of the
referee's own intellectual property is tied up with the reputation ofhis
field, and of those journals with which he is most closely associated.
His own purposes are served by helping to ensure that the creation
306 Social Aspects ofScientific Activity
ofproperty in his field is accomplished in an orderly manner, that the
shares in property are allotted fairly through acknowledgement of
previous work, and that the value of the property of all the mem-
bers of the group is not allowed to deteriorate through admission
of shoddy work. Thus, provided that he is protected from undue
pressures from the other producers, and given some direction and
control from above, the referee need be concerned only with the
protection of his group's intellectual property, to accomplish his
tasks adequately.
Unless the referees are indeed committed to something more than
the protection of property, control and direction from above is very
necessary for the prevention of a rapid decline in standards. As we
have seen, effective control on the product of such a skilled task as
scientific research, is possible only when those exercising the control
are themselves expert craftsmen. But then they will be working on
two sorts of tasks, involving the different private purposes associated
with production and control respectively. Those whose short-term
private purposes are served by the lowering of standards for the
easier production of intellectual property, will not be a separate and
inferior group; they include colleagues and proteges of the referees,
and the referees themselves. It is even possible, in small and special-
ized fields, for two good friends simultaneously to referee each
other's papers. In such an extreme case, it is necessary for the eti-
quette of refereeing to be adhered to very rigidly, if mutual con-
venience is not to dominate over the collective goal ofthe production
of genuine results. But here, unlike in the case of the professionals'
division of a complex task, the etiquette is not automatically en-
forced by the aggregate of self-interests of practitioners. And this
very subde method of social behaviour is very vulnerable to the sort
of modification at the fringes, by which no apparent damage is done
in the short run, but whose cumulative effects are quite destructive.
Hence, as we discussed earlier in the case of quality control, the
role of the leaders of the field is crucial. If their concern is no more
than the creation of intellectual property which can be cashed for
material and social benefits, then there are no internal barriers to the
rapid degeneration and corruption of a field at all levels. Aspirants
to leadership will inevitably have a complex mixture of motives,
involving both personal ambition and scientific zeal. They will take
the style of behaviour of their superiors and patrons as more signifi-
cant than their public pronouncements for shaping their own
 Ethics in Scientific Activity 307
behaviour and purposes. In a single generation, the style of leader-
ship can change from that of genuine scientists who are greatly
enjoying unexpected material benefits, to entrepreneurs in the
business of the production of results on conttaet. Hence if the
leaders are to maintain the health and vitality of their field, their
private purposes must be at least as refined as those of the leaders of
a profession who are committed to a genuine professional ethic.
The Social Context of'the Scientific Ethic'
The 'scientific ethic', analogous to a 'professional ethic' and the
ideology which necessarily serves as its basis, lacks the direct con-
tact with experience ofthe professional ethic, and is correspondingly
more sensitive to the social and cultural context of the work. For the
professional's task is defined by the needs of a particular client; and
the ultimate purposes of the task, in serving the group of clients, are
present every time the work is done. By contrast, the higher final
causes which condition the scientist's task are remote and diffuse.
The collective goal of scientific endeavour is the increase of know-
ledge; but it is impossible to identify a group in society whose
purposes are so directly served by the continued fulfilment of that
goal, that they can be said to determine it and all the final causes
down to the specification of the goals of a particular task. It is true
that the work and the achievements of science do perform various
functions outside scientific enquiry itself, through contributions to
technique, through the maintenance of the quality of high-level
teaching, and in enhancing the prestige of a local or national in-
stitution or community associated with it. But in the ttadition of
science and scholarship that we have inherited, these external
functions do not, and must not, provide ultimate purposes which
dominate the final causes whereby scientific work is directed and
controlled. In this tradition, the collective goals of scientific inquiry
are autonomous; this not only creates an inherent tension between
the community of science and the lay society which supports it, but
also deprives the ideology and ethic of science of a firm basis in
experience. Assuredly, the search for knowledge has long been
recognized in our civilization as one of the most noble of human
endeavours. But lasting success in this search is reserved for only the
very few; and no one in science can be sure that he will be other
than a humble participant in a task whose proper accomplishment
may require the work of generations yet unborn.
308 Social Aspects ofScientific Activity
The 'scientific ethic' to which the leaders of science must sub-
scribe if they are to do their work properly is thus so refined and
subtle as to rest ultimately on philosophy and faith. Francis Bacon
defined the attitudes which the ideal natural philosopher should
bring to his work: he should be 'sober, chaste and severe', and should
approach his task in spirit of humility and innocence. He was re-
acting against the various corruptions of knowledge which he saw in
his own time, principally the vacuous disputations of academics and
the boastful, spurious claims of the natural magicians and al-
chemists; and for him the reformation of natural philosophy was
intimately related to a true reformation of morals and religion. I I
Although his personal conception of the search for knowledge was
soon lost, his definition of its goals, 'to establish and extend the
power and dominion of the human race itself over the universe',
remained influential for generatioDS, for it cohered with the ex-
periences and values ofnatural philosophers in many ways. 12 Among
the English until nearly a century ago, the study of God's creation
was (as Bacon himself believed it should be) parallel to the study of
His word, and it was widely believed that this study could not but
increase reverence for the Creator.13 Elsewhere this 'natural theo-
logy' was not so sttong, but the close relation of scientific inquiry to
metaphysical concerns, especially among the greatest men as in
Germany, provided the work with an ultimate purpose harmoniz-
ing with its immediate goals. Also, the general social benefits, in the
way of material progress, thought to be largely a result of the ap-
plications of science, provided a justification of the work and an
added component of human value; although (as we have seen in the
II This interpretation of Bacon is developed in Benjamin Farrington, TIte PJaIDSOI1J,
0/ Fraas BMtnI.
Or,,,,,,.
la This formulation of the goals ofDatural philosophy is found in the NtnJIIIII
I, Aphorism 1291 Woris, voL 4t p. 114- Bacon's essential message bas been variously seen
u 'induetiYism', as 'power over nature', and as 'the sacred mission of scientific inquiry'.
The fint interpretation was cbaracteristic of the nineteenth-centary British; the second
has a history from the Enlightenment through to the Manists, and the third is recent.
The writinp of Benjamin Farrington show a traDsition from the second to the third
interpretation; see his Fraas BMtnI, PllilDsol_ of ltUltutrW Stinlee (New York,
AbeIarcl-Schuman, 1947) and TIte PllilDsol1Jy ofFraas BMtm.
II The concluding paragraph of Newton, 0lti,is, clearly expresses this higher function
of Datural knowledge; and this tradition continued until well into the nineteenth century.
For the continuity between Newton's fundamental Datural philosophy and his ~
Peal CODCCI'DI, see J. E. McGuire and P. M. Rattmsi, 'Newton and the Pipes of Pan',
Noles Mil Re,ortls oflite Roy," StKiny, 21 (1g66), 108-43, and the later work of J. E.
McGuire, such u 'Force, Active Principles, and Newton's Invisible Realm', -AJUiz,
15 (1g68), 154-208.
 Ethics in Scientific Activity 309
case of Helmholtz) the greatest spokesman of science were careful
to warn that this external function should be kept out ofthe hierarchy
of final causes in science.
These very diverse strands of ideology could serve as an intel-
lectual basis for a refined scientific ethics, because the social position
ofscientists, either by birth or acculturation, provided a natural basis
in experience for such an ethic. For, until quite recently, one could
generally not engage in free scientific inquiry if one had to earn a
living. Even where there was a developed institutional structure, as
in Germany, a 'career' in science involved years or decades of
living on independent means or on nearly none. Hence the vocation
of natural philosophy required, for all but brilliant recruits from the
lower orders, the background of a gentleman. 14 Those of this class
who were concerned with ordinary comforts or social position
would not devote themselves to philosophy; and those who made
the sacrifice would do so in the pursuit of more refined satisfactions.
Among them in particular, the code ofa gendeman, in which a man's
most valuable property is his good name, would be powerful. In
scholarship, a good name is achieved by talent and protected by
integrity; and an overriding commitment to the quality and trust-
worthiness of work with which one is personally associated in any
way is the essence of a genuine scientific ethic. IS
14 Thus Helmholtz, in his autobiognphy: 'And now I was to go to the university.
Physics was at that time (in the late 18305) looked upon as an art in which a living cou1d
not be made. My parents were compelled to be very economical, and my fatherexplained
to me that he knew of no other way of helping me to the study of Physics, than by taking
up the study of medicine into the bargain' (p. 274). Popu/M Lertw,s tnJ Seimtiji& Subj,rls,
second series. The weakness of the position of the PriVflttlountm in the German
universities of the nineteenth century, which eventually became intolerable when the
professoriate did not increase in proportion to student numbcn and research work, is
discussed in J. Ben-David, 'The Universities and the Growth ofScience in Germany and
the United States', Minerva, 7 (1968-"9), 1-35. He finds the cause of the success of
American universities in their structure, while it seems more likely to have been in their
growth-rate; and his chosen example of statistics as a 'growth-point' is ill-suited to his
argument, since it flourished in England 10Dg before it did in America.
15 A particularly stem and perceptive analysis of the scientific ethic was given in the
nineteenth century by Fresenius, the founder of analytical chemistry: 'Knowledge and
ability must be combined with ambition as well as with a sense of honesty and a severe
conscience. Every analyst occasionally has doubts about the accuracy of his results, and
also there are times when he knows his resu1ts to be incorrect. Sometimes a few drops of
the solution were spilt, or some other slight mistake made. In those cases it requires a
stroDg conscience to repeat the analysis and not to make a rough estimate of the loss or
apply a correction. Anyone not having sufficient wiD-power to do this is unsuited to
analysis no matter how great his technical ability or knowledge. A chemist who would
not take an oath guaranteeing the authenticity, as well as the accuracy of his work,
II-S.K.S.P.
310 Social Aspects ofScientific Activity
In the conditions of the industrialized science of the present
period, all these bases for the maintel1ance of a scientific ethic have
been eroded: philosophy and religion have become irrelevant; the
applications of science are not a matter of simple benefit; and one
can earn a living and indeed become rich in a career in science. 16 The
great pronouncements of the noble ideology of science of the earlier
period would sound false or hypocritical in this new context. These
changes produce many difficulties in the preservation of an enlight-
ened ethical code in science. I? The leaders themselves are deprived
of a secure basis in their personal culture and experience for the
ideology and ethics ofautonomous, dedicated scientific inquiry to be
natural. Should they parade such old-fashioned beliefs as arguments
for adherence to the demanding ethics of science, they would risk
isolation from a large section ofthe community they lead. Moreover,
even these leaders do not exist as superiors in a formal hierarchy of
direction and control; the governing of their community can be
accomplished only with the active participation of a large part of it,
should never publish his results, for if he were to do so, then the result would be detti-
mental, not only to himself, but to the whole science! Q!Jotation from F. Szabadvary,
HisltWy of ,Au/yti,,,1 Cllemistry (pergamon Press, 1966), p. 176. I am indebted to Dr.
W. H. Brock for this reference. It would be naive to believe that all gendeman have
always been gentlemen. In his discussion of observations, in ch. 5 of Refle'tions 0" the
D"li. of Same, i" E"gland (London, 1830), Charles Babbage devoted a section to
frauds; he defined hoaxing, forging, and trimming, and then, with fine irony, de-
saibed the varieties of 'cooking'; see pp. 174-83.
16 Another danger, which until very recendy was so remote as to be inconceivable, is
that scientists might resort to the law of libel to protect their results from published
criticism. On this, see]. R. Lewis, J. S. Gray, L. W. Pollock, and P. G. Moore, 'Can a
Scientific Article be Libellous?' (letter), Nat"e, 225 (1970), 1081. The Master of the
Rolls, quoted in TIN Times, was reported as commenting 'it would be a sorry day if
scientists were to be deterred from publishing their findings for fear of libel actions. So
long as they refrained from personal attacks they should be free to criticise the findings
and techniques of others. It was in the interest of truth itself. Otherwise, no scientific
journal would be safe.' As the matter is still sub jrulU" the merits of the case cannot
be here considered; but, since the original report and the criticism appeared in a
'scientific' periodical (the BritisIJ M,tli&"l JoUf1UjI), the outcome will naturally be
awaited with anxious interest by those concerned to protect genuine scientific criticism•
•, An echo of the old ethical code of science is found in the review, by Hedwig and
Max Born, ofT1Je S&imtist SpenJiu,s:." ,AfIIlIology ofPartly-Baietl Idlas, ed. I.]. Good
(Basic Boob, Inc., New York, 1963), in BlIllet;" oftIN Atomi, Sdmtists, 19 (May 1963),
30-3. The review starts: 'The publication of [this book] is nothing less than a aime
against the ethical code, unwritten but vital, of the community of scientists.' Later:
'UllQ)IDpromising, indefatigable pursuit oftruth, then, is the hallmark that distinguishes
the scientist from the charlatan. It constitutes the indispensable ethic of science.' They
object particularly sttongly to the making of priority claims on the basis of half-baked
spec:ulatioDs. The review concludes with a list of blunders and absurdities contained in
the book.
 Ethics;" Scientific Activity 311

and the willing acquiescence of the rest. Should morale decline

among the ordinary members of a scientific community, then it
becomes impossible, either by sanctions or rewards, to restore it
through actions taken at the top. And without good morale, what
good work is done will fairly soon be driven out by the bad. There
will inevitably be some leaders who will strive only for instant
prestige; the referees associated with them will take the hint and
authenticate any property which will enhance or share in that
prestige; and ordinary scientists will be under pressure to achieve
their private purposes most cheaply by identifying the current
fashions and producing passable results in their image. If this
degeneration occurs in isolated fields, it can still be checked through
the system of controls across fields. But if corruption spreads to this
highest level as well, then whole areas ofscience can become gigantic
confidence-games, producing pseudo-property at a feverish pace,
and resembling a stock exchange in a bull market rather than a
collective endeavour on behalf of the highest human goals.

Ethics in Sc;ence: Old Pronouncements and New Problems

The most eloquent pronouncements of the ethical basis ofscience
came at the end of the academic period, when their basis in social
reality was on the point ofextinction. In retrospect, they can be seen
to suffer from the common fallacy ofdefining a social activity in terms
of the idealized attributes of its perfect state. As materials for an
analysis of the ordinary practice of scientific research, they have
been recognized as gravely defective. But when they are recognized
for what they really were, statements of a credo and warnings of
danger, their true worth can be realized. They need not apply in
detail or even in principle to every scientific worker who publishes
a paper; but if the spirit embodied in their separate insights is lost
to the leaders of science and to those who aspire to that position,
then the outlook for science is grim.
The three most influential spokesmen for the ethical commitment
of science in the 1930S approached the problem from different and
sometimes contradictory points of view. The physical chemist
Michael Polanyi concentrated on the intense personal experience of
scientific inquiry, which was necessarily undertaken by a community
offree and dedicated men. The philosopher Karl Popper saw science
as the embodiment of intellectual honesty, realized through the
principle and practice of criticism. And the sociologist Robert K.
312 Social Aspects ofScientific Activity
Merton saw in the 'ethos' of science the realization of the highest
standards of civilized human behaviour. IS All three identified
dangers to science: for Polanyi they lay in bureaucratic direction
and control; for Popper they lay in the dishonesty latent in the
denial of criticism; and for Merton they lay in the inherent conflict
between the norms of co-operative scientific endeavour and those of
lay society and the State. The paths which their thoughts took in
later years do not concern us here; nor need we attempt a detailed
assessment of their conclusions. It is, however, worth recalling that
they were engaged, each in his own way, in a complex debate
centring around the Marxist view of society and of science. 10 The
Marxist writers, of whom J. D. Bernal was the most influential,
were primarily concerned with the external, social functions of
science. They believed that science had an obligation to contribute
in a conscious and planned way to the benefit of mankind; and also
that this was being done, with no loss to the integrity and freedom of
science, in the Soviet Union. 20
Looking back from the other side of the watershed of the Second
World War and the subsequent industtialization of science, it is
easy to see that the proponents of an idealistic ethic were unrealistic
about the attitudes necessary for the ordinary practice of science to
be successful in the short run; and that the Marxists were grossly
ignorant of the problems of science under Soviet conditions. But

jectures and Refutations', ch. I of Ctmj,ttur'$ IJ"

II For Ie. R. Popper, see the revealing autobiographical account in 'Science: C0n-
Re{utIJtiOflS (Roudedge, London,
1963). He describes how, in 1919, he came to see thatlmulo-sei"", (astrology, Freud's
psychoanalysis, Adler's psychology, and Marxism) survives by seeking (and finding)
confirmations, while real science (Newton, Einstein) advances by making testable
conjectures. For Polanyi, see TIte lAfit ofLilJert;y (Routledge, London, 1951), for example
p. 51: 'The ooherence of science must be regarded as an expression of the common
rootedness of scientists in the same spiritual reality•••• Only then are the conditions for
the spontaneous c:o-ordination of scientists properly observed.' For R. Merton, see

(1943); reprinted in SotiIIl TltetW;y IJ"

'Science and the Social Order' (1938) and 'Science and Democratic Social Structure'
soeW Strwh6, (The Free Press, Glencoe,
Collier and MacMillan; revised edition, 1957),537-61. The later paper gives the classic
formulation of the 'ethics of science' with the four 'BonDS' of 'universalism', 'com-
munism', 'disinterestedness', and 'orpnized scepticism'.
19 Popper bad adhered to Marxism, to some extent, before his philosophicalconversion;
but he later devoted much time to its refutation. Polanyi was amsistendy anti-Marxist
and antkoUectivist. Merton took Marxism seriously and did a study of seventeenth-
century science from a 'utilitarian' point of view, parallel to that in which he related
science and Puritanism. See SotitIl TltetW;y, cbs. 18 and 19. It is tempting to speculate
whether the 'norms' of science, particularly that of 'communism' resulted from an
attempt to reconcile Weber and Marx.
ao J. D. Bcmal, T1Je SotitIl FlIII&titm ofSd""e (Routledge, London, 1 939).
 Et!zics in Scientific Activity 313
each of these approaches reminds us of a necessary component of a
scientific ethics appropriate to the conditions of the present and the
near future. The idealistic approach involved the assumption of the
excellence of scientific inquiry as a permanent reality, and from this
derived an enlightened ethical commitment of scientists as a logical
consequence. My argument has used the same components, but has
changed the mode ofcausality. I make the more realistic assumption
that the health and vitality of scientific inquiry are not guaranteed,
either by the objects of inquiry or by the social aspects of the work.
Hence, unless there is an effective ethic, even more refined than a
'professional ethic', this very delicate and sensitive work will not long
continue to be well governed or well performed. Yet it is unlikely
that the ideological basis for such a scientific ethic can continue to
be found in religion and philosophy, nor its practical roots in the
refinement of the code of conduct of an elite class. Therefore the
ethical basis of future excellence in science must lie in some other
ideals and experience; perhaps in a humanitarian commitment,
necessarily interpreted in a much more sophisticated fashion than
ever before.
 Part IV
SCIENCE IN THE MODERN WORLD
 INTRODUCTION

Up to this point, our discussion has been restricted, for the sake of
simplicity, to 'scientific' problems, where the goal of the work is the
establishment of new properties of the objects of inquiry, and its
ultimate function is the ·achievement of knowledge in its field. For
an historical or philosophical study ofscience in the period preceding
its industrialization, such a restriction is natural; then, it was only
occasionally, and in certain specialized fields, that inquiry of this
sort could be applied directly to the mastery of the natural world.
But with the interpenetration of science and industry, this tradi-
tional separation has come to an end. Indeed, most of the tasks
undertaken by the contemporary corps of 'scientists' have very
different functions, and even different goals as well. The emergence
of this new sort ofwork has led to a variety of new descriptive terms,
which reflect the confusion over its nature and its relations with
science of the traditional sort. The term 'science' itself has stretched
to include technology, and even any sort of inquiry whose methods
are modelled on those of the experimental and mathematical natural
sciences. This multiplicity of meanings is all encased in the single
plausible tide of a projected new field of inquiry: 'the science of
science'. Since these new sorts and extensions of 'science' are those
which present the most severe problems in their relations with the
social and natural environment, we must try to achieve some under-
standing of their distinctive character. I
From one point of view, all the new sorts of 'science' are entided
to that honorific appellation: their work consists of problem-solving
on intellectually constructed objects; and there is no clear demarca-
tion between the methods used in traditional pure science and those
in the most speculative ofsocial technologies. We can, however, make
a useful distinction among the various sorts of problems by means of
I The 'philosophy of technology' is a subject which has attracted relatively little
attention in recent years. The record of a symposium on the subject is published in
Teclmolov "nd Cul'ure, 7 (1g66), 301-40. In this collection of papers, that by Mario
Bunge, 'Technology as Applied Science', bears the closest relation to the problems
discussed here.
318 Science in the Modern World
the classification of final causes involved in a task, that I developed
in connection with the problem of quality control in science. We
recall that the task itself has a goal, which is conditioned more or
less strictly by the function which will be performed by the result
ofthe accomplished task; and this in turn is governed by the ultimate
human pwposes which are expected to be served by the performance
of that function. In the previous discussion, we were concerned with
the need for a harmony between these 'objective' final causes and the
'subjective' ones brought to the task by the scientist: his own pur-
poses, in what he is trying to achieve for himself in the work; and
his motives, which bring him there and set his purposes within the
constraints of the task. For the present problem, we shall con-
centrate mainly on the 'objective' final causes, and by their means
distinguish between problems capable of description as 'scientific',
'technical', and 'practical'.
In each case, we consider the task itself as the investigation of the
problem; the 'goal' of the task is the solution of the problem. We
then ask, to what extent are 'functions' and 'purposes' involved in
the specification of the goals of this sort of task. In the case of the
'scientific' problems that we have been discussing up to now, these
higher final causes are not so influential. The function to be per-
formed by the solved scientific problem, that is the contribution of
new results for the advancement of the field, conditions the work
only in a general way, through the controlling judgements of
adequacy and value. The investigation of the problem may change
course in mid-stream, if a more promising solution appears possible;
and then the solved problem will perform a different function, to
no one's regret. The ultimate pwposes to be served by the
scientific problem are quite remote, diffuse, and unpredictable in
detail; we recall the words of Helmholtz on the radical separation
between the applications of science and the goals of the scientific
endeavour.
By contrast, we describe as 'technical' problems those where the
function to be performed specifies the problem itself. The goal of the
task is fulfilled, and the problem 'solved', if and only if that function
can be adequately performed. A design team that produces a device
that performs a function different from the one assigned, will not
usually be congratulated on its initiative and independence. On
the other hand, in a technical problem the work is conditioned only
in a general way by the purposes to be served. Thus, in a com-
 Introduction 319
mercial product, what counts is whether a sufficient number of
people expect the product to serve their purposes, and so buy it;
exactly why they like it does not matter except as a guide to the
advertising.
Finally, there is a class of 'practical' problems in which the goal
of the task, in principle, is the serving or achievement of some
human purpose. The problem is brought into being by the recog-
nition of a problem-situation, that some aspect of human welfare
should be improved. Such problems involve intellectually con-
sttucted objects rather than ordinary 'common-sense', as we see
from programmes and debates on such issues as 'education', 'poverty'
and 'health'. We shall see that large-scale practical problems give
rise to subsidiary technical problems, and perhaps to scientific
problems as well. The fulfilment of its goals in a multitude of in-
dividual cases will usually require a large-scale project, with a
hierarchy of decision and control and a division of labour, analogous
to the projects required for the complete solution of a large-scale
technical problem. 2
By means of this distinction we can consider the similarities and
differences among the three basic sorts of problem, as well as the
various sorts of difficulties to which they are prone, individually and
when they are related in practice. We shall see that technical and
practical problems each have their cycles of investigation, and are
subject to controlling judgements analogous to those of scientific
problems. Even in the case of technical problems, there is no auto-
matic mechanism enforcing quality control on solutions, and one
can speak of a degeneration and even corruption in technology as in
science. Practical problems, and the large-scale practical projects
required for their execution, are by their nature subject to the most
severe hazards and pitfalls. An understanding of these is most
urgendy necessary; for the increasing complexity and sophistication
of our society throws up a rapidly growing set of practical problems.
Even problems previously considered as purely 'technical' are
3 My c1assification of scientific, technical and pnctical problems is closely parallel to
Aristode's classifications of 'theoretical', 'productive', and 'practical' knowledge. See
the Ni'01NU1Ie." Ethi's, vi, 3-5. 1139bI4-114Ias, Met.physits, E.I.1025b, and Sir
David Ross's commentary on the latter text in his edition (Clarendon Press, Oxford,
1924). Aristode distinguishes between 'making' and 'acting', and observes that the
latter is not governed by any end outside itself. This corresponds to my distinction
between 'performing a function' and 'achieving a purpose' for the goals of technical
and practical problems respectively; the function to be performed by a device is always
governed by some ultimate purpose or purposes.
320 S,imte in the Modern World
increasingly being debated in their 'practical' aspects under the
slogans of 'environment', 'pollution', and 'amenity'.
To achieve this understanding, we must appreciate the con-
sequences of the ineffectiveness or immaturity of the disciplines
(usually those concerning man and society) on which the investiga-
tion of practical problems depends. The analysis of the state of
immaturity is not a pleasant task; the condition itself is generally
regarded as at best a misfortune to be remedied quickly, or at worst
as a scandal to be concealed. But without a knowledge of the deep
and systematic differences in the practice of inquiry, and in the sort
of results obtainable, between immature and mature disciplines, we
will have no way of controlling the harmful effects (to themselves
and to society) of ineffective fields grown beyond their appropriate
size and prestige. The analysis of the credence given to the claims of
patendy ineffective disciplines leads us into a discussion of the
category of 'folk-science', which applies equally well to the most
sophisticated academic exercise as to peasant superstitions. With
this battery of concepts we can return to a discussion of the hazards
and pitfalls of practical projects; and see that the wreckage of so
many of the great schemes for human betterment is, sadly, only the
natural state of affairs, at least so long as our conception of 'science'
is that inherited from Galileo and Descartes.
 12

TECHNICAL PROBLEMS

As a result of the industrialization of science, the traditional lines of

demarcation between 'science' and 'technology' are blurred and
confused. One cannot distinguish the two sorts of problems by
the formal qualifications or the formal tide of employment of the
research worker, nor even by the name and stated character of the
larger project of which a particular investigation forms a part.
Indeed, a single problem may have a mixed function, and so be
both 'scientific' and 'technological' in varying degrees to the
different parties involved in its planning, investigation, and applica-
tion. If the real differences between 'science' and 'technology' are
to be appreciated, and their interactions controlled for mutual
benefit, there must be a clear understanding of the differences
between the problems characteristic of each sort of work.
We can introduce the class of problems described here as 'tech-
nical' by recalling the cases where such problems have been mentioned
in connexion with scientific inquiry earlier in this account. First,
there are those problems whose goal is the establishment of new
properties of the objects of inquiry, but where the function is
assigned, externally and precisely. This is where certain information
is required for the successful solution of a problem, and the pro-
vision of this information constitutes a subsidiary scientific problem
in the work. The external, determining problem may be a scientific
one, and the subsidiary problem may be investigated by the
scientist himself, or by an assistant. But in this situation, the use to
which the information will be put, its function, constitutes the
major component of the value of the subsidiary problem, and it
determines the standards of adequacy by which its solution will be
assessed. The investigation of the subsidiary problem is then
governed by criteria of quality which are not at all the same as those
for a scientific problem; and even though the result has the same
322 Science in the Modern World
form as that of a scientific problem, we gave it a different name,
technical problem.
A more pure case of a technical problem is also encountered in
scientific research: the making oftools. Here the function determines
the problem, but the solution of the problem, instead of being the
conclusion to an argument, is the creation of a device, the means to
the performance ofthe function. Sometimes such work involves only
the application of simple craft skills, even if the objects themselves
are defined in terms of scientific categories; this can be the case, for
example, when one assembles standard electronic components to
create a particular measuring instrument. Then the work is best
described as a technical task, rather than as a technical problem. The
latter tide is better reserved for the work of those scientists who
specialize in tool-making, where the sophistication of the tools is so
great that the work involves drawing conclusions to arguments con-
cerning the properties of intellectually constructed classes of things
and events. Technical problems, in this sense, are not restricted to
scientific tool-making; this is the sort of work required whenever
there is a function to be performed, and its new means cannot be
created by craft knowledge alone.
The penetration of industry by science in the present period is
possible because the set of soluble technical problems is large, in-
creasing, and continuously growing in importance. In early modern
times, the only such problems were in the descriptive-mathematical
sciences: navigation, surveying, and their allied fields, and arithme-
tic. A significant advance did not come until after the industrial
revolution; although in the eighteenth century there had been a
development of the application of an experimental approach to
practical tasks (as in Smeaton's analysis of waterwheels), I and
also some innovation which used established scientific results as
information. But until the later nineteenth century, the intellectual
objects of scientific inquiry were generally so removed from the
physical objects of production, that attempts at analysis and
innovation in their terms were generally fruidess. 2 The effective
I Dr. D. S. L. Cardwell and D. A. }. Pacey, of Manchester, have shown the sig-
nificance for science as well as for engineering, ofSmeaton's work; see D. S. L. Cardwell,
'The Academic Study of the History of Technology', History of Sdmet, 7 (1g68),
119-20.
3 I am indebted to Professor A. Kauffeldt, of Magdeburg, for a systematic develop-
ment of this point (private correspondence). He observes that the impotence of theo-
retical science, and its resulting irrelevance to the immediate needs of society, were
 Technical Problems 323
penetration of science into industry first came in those areas where
the products themselves were pure and artificial, as in synthetic
dyestuffs and electricity; and even to this day, it is more difficult
to solve technical problems in industries whose materials and pro-
cesses are largely traditional. 3 But the largest and most rapidly
growing sections of industry in the advanced economies are those
whose basis is scientific, and it is in their 'R. and D.' sections that
the solution of technical problems has become a major industry in
itself: We may consider 'technology' as the activity of investigating
such technical problems, as 'science' is the activity of investigating
scientific problems; and in this way identify the similarities and
differences between the two sorts of work.
The Cycle ofInvestigation
By analogy with a scientific problem, we may define a technical
problem as a statement of a function to be performed, whose means
are to be established as the conclusion of an argument, with a plan
for its accomplishment. The similarity to scientific problems is
brought out by the mention ofthe conclusion of an argument, which
implies a discussion ofclasses ofintellectually constructed things and
events. However, a technical problem will generally have less freedom
to evolve than a scientific problem; to the extent that the func-
tion is externally assigned, and is not the speculation of an inde-
pendent inventor, it cannot easily be altered if some unexpected
possibility or difficulty appears in the course of the research. Also,
the function will generally be specified in a more detailed fashion,
each of its aspects being assigned some standards of adequacy of
performance. This further restricts the freedom of the problem to
evolve, and also makes its cycle of investigation significandy
different from that of a scientific problem. 4
useful in providing it with the autonomy in which to develop along lines which were
ultimately very fruitful.
3 'No newcomer to textile research can help but be impressed with the degree of
optimisation of the processes that has been achieved by trial and error methods extend-
ing over a long time. The same cannot be said for the conventional theories that have
gone with these developments. One of the earliest needs of the textile engineer who
wishes to modify existing processes is to disentangle the true purpose from the sup-
posed purpose of many textile machines.' P. Grosberg, 'The Changing Role of the
Textile Engineer', Inaugural Lecture (Leeds University Press, 1965), p. I I •
.. This schematic account necessarily abstracts from many of the aspects of getting
a technical problem under way, in particular, those of decision which we have discussed
in connection with science. On this, see E. B. Roberts, 'Entrepreneurship and Tech-
nology', in Factors ill the T,,,,,sf~ ofTttlJlJOloV, cds. W. H. Gruber and D. G. Marquis
 324 Science in the Modern World
As an example, we might consider the creation of a new model of
aeroplane. 5 The project is undertaken only when someone believes
that a certain function could be better performed than at present.
This is not merely 'carrying passengers by air'; it will concern a
class of traffic defined by load in passengers and freight, distances
travelled between stops, available landing and maintenance facilities,
climatic conditions, and required safety and reliability of service,
all of these with estimates of averages and extremes, and translated
into a variety of capital and running costs and returns. Until some-
one conceives of a model of aeroplane that can adequately perform
the particular function just described, we may say that a technical
problem-situation exists, but not a technical problem. The problem
comes to be when a possible model is tentatively specified. This
specification will not be in terms of the function itself, but of the
operating characteristics of the proposed model. This distinction is
important, for it is the first of the conceptual shifts that occur when
a technical problem is being investigated. For example, when
'function' is being discussed, landing facilities will be defined by
length, orientation, and quality of runways, provision of guidance
systems of different sorts, and relevant topographical and climatic
features. The landing properties of the aeroplane, on the other
hand, are defined by certain parameters of its own behaviour, in-
cluding landing and stalling speed, angles of ascent and descent,
taxiing distances, along with properties of its installed guidance
system, undercarriage, and other relevant features. We should notice
that both the function and the operating characteristics are defined
in terms of scientific categories, rather than craft knowledge or
informed common sense. The numbers by which they are described
are based on tests using sophisticated equipment and a variety of
theoretical assumptions, under specialized conditions. 6 Thus even
(M.I.T. Press, 1969), pp. 219-37. The same volume contains a review by D. }. de S.
Price on the work on patterns of citation and communication in science and in tech-
nology: 'The Structures of Production in Science and Technology', pp. 91-104-
5 A very clear account of the cycle of a technical problem, applied to a new model of
an aeroplane is given by D. Norman and J. Britten in 'The Islander Project', AtlVllnet
(University of Manchester Institute of Science and Technology), No.6 (April 196g),
Pp·44-9·
6 Working engineers are necessarily familiar with the theory-laden and imprecise
character of specified operating characteristics in their own field; the general public
learns of them only incidentally. Thus, the 'True Boiling Point' ofa pettoleum distillate,
the only way to 'establish its basic marketing specification', 'leaves a lot to be desired'.
It requires a lengthy process for its determination, 'the results are useless in conttolling
a refinery', and 'the method is not very reproducible [si"] especially as it applies to the
 Technical Prohlems 325
in this early phase of the work, there is a transition from one set of
intellectually constructed objects to another; and an argument is
required to establish the conclusion that a device with these par-
ticular operating characteristics will perform the assigned function
to the given standards of adequacy. In this example, the argument
will be devoted to showing that an aeroplane with this particular
landing behaviour will be able to make reliable and safe use of the
anticipated landing facilities under an anticipated range ofconditions.
We should notice that suth an argument involves approximations
and invokes considerations ofmore or less, and invokes probabilities,
as an essential part of its content.
We notice that the full specification of the 'function' includes
rather more than the use to which the device will be put; we also
find relevant aspects of the conditions under which it will operate,
and of its maintenance in operation. Although this extension may
appear to stretch the concept beyond its original meanings, it is
easy to see that these additional components of the function must
be specified in order that the device as designed will do its job. Also,
each of these components of the function must be translated into
appropriate operating characteristics. As in scientific inquiry, pit-
falls can be encountered at every stage in the work; and in the
process we have described so far, two sorts are possible. First, there
may be faults in the translation from the specified function to the
operating characteristics; so that even if the device is constructed to
meet these, it may yet fail to perform its assigned function. This pit-
fall is most likely to be encountered when a concentration on the
most important aspects of the function leads to the neglect of the
implications of something apparently minor. This can happen most
easily when the original set of operating characteristics proves to
be unrealizable, and the task of re-designing is undertaken under
pressure of time. The revised set of operating characteristics, tested
against the main aspects of the function, may appear to be adequate,
but it is possible either that some of the characteristics may be
mutually incompatible, or that some aspect of the function, which
was satisfied as a matter of routine in the first conception, is satisfied
no longer. A more serious and fundamental pitfall is in the incorrect
specification of the function. This occurs when there is insufficient
craft knowledge of the actual conditions in which device will
initial and final boiling points.' Hewlett-Packard offers this infonnation, as background
to its improvement upon it. See advertisement in S,imtl, 167 (1970), 1169-70.
 326 Science in the Modern World
operate, so that the categories in which the function is specified fail
to include all the relevant aspects of the real situation. 7 This par-
ticular sort of pitfall is most likely when radical innovation is being
attempted; it is impossible to be sure what the working conditions
will be like until the device is actually working, and it is impractical
to over-design in every conceivable aspect for safety. One mark of
the great engineer is the ability to penetrate into the unknown in
this way ,rith success.
Once the operating characteristics are specified, the cycle of
investigation of a technical problem proceeds through several
further phases, each of them involving a translation between con-
ceptual objects, an argument, criteria ofadequacy, and characteristic
pitfalls. The next phase is design, providing specifications of the
structures, motions and actions by which the operating character-
istics can be realized, as well as the materials in which the structures
and processes are embodied. From the design, detailed plans are
drawn up, providing the basis for the construction of the physical
object itself. This end of the cycle can vary gready in its com-
plexity, depending on the sophistication of the device and of the
process of production, and on the quantity to be produced. There
may be a series ofprototypes, produced under conditions increasingly
7 The influence of the theory in whose terms a device is conceived on the specification
v.
of the function is shown well by an example of Ronchi, Optics, the Sden&, of Vision.
His experience of the design of nnge-finders is that they are designed to the highest
possible accuracy, by the installation of extra lenses and plates. But this refinement is
irrelevant when the instrument is used under conditions of low illumination, when all
the extra optical parts cause a loss of energy to below the threshold of visibility. This
neglect of the conditions of use stems from the neglect of the eye of the observer in
physical optics,which he discusses at length (p. 283). The imposition of quite arbitrary
aiteria of adequacy in the performance of a function could be seen in certain charac-
teristics of the Systemc International of measurements, as first announced in the
United Kingdom. With a fine rationalist zeal, the designers of the system determined
to reduce the number of ofticial prefixes denoting magnitude, so that the memory of
users would not be overburdened. One result of this was the banishment of the 'deci-'
and 'c:enti-', denoting tenth and hundredth respectively. Then all new modules in
industry were to be given in millimetres; this extended to builders' bricks. In this the
designen of the Systemc neglected the well-known fact that the recording of measure-
ments is onerous, liable to blunders, and also misleading in its implied precision, unless
the common magnitudes are expressed as small multiples of submultiples of an appro-
priate unit. One group of scientists were able to force a violation of the 'thousands' rule
for derived units, and achieved an official unit for pressure, very close to one atmos-
sphere, which is 105 times the standard unit. On the other hand, the common run of
humanity was expected to put up with measuring lengths either in decimal parts of a
metre, or in large multiples of the millimetre. It appears from casual indications that
recently this hyper-rigorous approach has quiedy been relaxed.
 Technical Problems 327
resembling those of full-scale production; and the creation of the
means for the production may itself involve technical problems of
great difficulty. The full cycle, for a commercial product, is not
complete without regular production, followed up by promotion,
sales and service. As the cycle advances to its latter phases, the work
thus changes from that of investigating a technical problem, to what
we may call accomplishiqg a technical project. The scientific and
technical component diminishes, and is replaced by routine and
craft work, social organization, and (for most devices) commercial
considerations. In one sense these are an integral part of the cycle,
for it is only when the device brings in financial returns that the
ultimate purposes for its creation can be achieved; and its success
against competition is an important part of the test of its adequacy.

Technical and Scientific Problems: Differences and Similarities

The process of re-cycling during the investigation of a technical
problem will have important differences from that of a scientific
problem. Once a scientific problem has been set, the materials
produced in an earlier phase serve as the basis for those of later
phases; as 'causes' in the Aristotelian sense, they are material, and
only when they are inadequate to their function is there a need for
a revision of the plan of inquiry. Even then, the revision may be
either in the direction of improving the materials, or in altering the
goal of the task in the light of the experience with these materials.
Of course, even in science such a re-cycling does not always succeed;
the materials may turn out to be useless for the establishment of any
worthwhile conclusion. In technical problems, the relation of the
earlier phases to the later ones is more that of final causes; the as-
signed function delimits the possible operating characteristics,
which in turn delimit the design, and so on. Re-cycling becomes
necessary when it appears that the requirements set by the results of
an earlier phase of the work cannot be met, either from technical
incompatibility or (more commonly) from the preassigned limits on
costs. Then every significant change in plan has effects running
right back up to the definition of the problem itself, and hence the
existence of the technical project. Thus a difficulty encountered at
the prototype stage may require changes in design which alter the
operating characteristics and so yield a modified device which does
not perform the preassigned function as well as had been expected.
Hence while the problem lacks the same freedom to evolve as a
328 Scim&e in tile Modem World
scientific problem has, the constraints imposed on the solution of
problems at the later phases present challenges, and require control
and co-ordination, to a degree which is not present in science.
In spite of this interrelation of phases, the difference between the
earlier and later tasks enable them to be considered as distinct, in
their organization as well as conceptually. Especially in the most
sophisticated industries, the phase of'research and development' can
involve exploratory work where the preassigned function is not
always narrowly defined. 8 In the case of such work, we can speak of
the investigation of technical problems in the strict sense, inde-
pendendy of the technical projects of which they may form the
earlier phase. Because of this natural division, it is possible for a
technical problem to be successful in one sense while failing in
another; the 'creation of a device' does not succeed or fail as an
indivisible whole, as does the investigation of a scientific problem.
This division is not a new one; the history of technology accords
posthumous honour to men for the greamess of their conceptions,
in spite of the occasional failure of their creations at the practical or
even technical levels.
Although the ultimate purposes and even the immediate functions
of technical problems are so different from those of scientific prob-
lems, the goals of individual tasks can be similar, and the methods
of work nearly identical. In both cases, the solution of problems
involves argument, based on controlled evidence, and cast in terms
of conceptual objects. Because of this, someone with the training
and oudook of a technician will be incapable of accomplishing the
tasks; while a scientist, once adjusted to the new sorts of problems,
will work as well as someone trained up to technology. In fact, the
differences between scientific problems and technical problems
become prominent only when the work is considered in the large.
In routine, detailed work, the two sorts of problems overlap con-
I In his review ofR L. NiebUI'J, /" 1M NIUM ofStimte (Qpadrangle Boob, Chicago,
19(6), in the Billie';" oflile AltnIIie StinJlists, 22 (November 19(6), 22-3, Bernard I.
SpiDnd distiDguisheI two stages of 'It. and De': 'the production of a worbble device
iUUlb'atiDg a principle, and the production of a device using the principle for specific
pIUs'. In the language of this analysis, one first sees whether a particular idea can be
rea1izcd in a model whose operational cbancteristics are absolutely minimal; and one
afterwards investiptcs the possibility of improving the design so that the device can
perform some recognized function. He argues that some projects labelled as 'boon-
dogies', • the 'Dyuuaur' and the nuclear-powered aeroplanes, were actually serious
projects; and for support he cita the nuclear submarine, which originated in just this
way.
 Technical Problems 329
siderably, and a given problem may be capable of being considered
in both classes simultaneously. The social distinction between
science and technology, which is of great practical importance in
England, derives not from any essential difference between the two
sorts of work, but from a combination of inherited social attitudes
(which are not so common on the Continent), together with aca-
demic prejudices import~d from nineteenth-century Germany. But
the idea that someone trained as a 'scientist' is thereby rendered less
willing or less capable at working on technical problems has little
relation to reality at the present time.
Yet there are certain deep differences between the two sorts of
work; so that for the preservation of the health of both they must
be kept distinct while yet in contact. It is well known that a scientific
result may give rise to a technical problem, when it is realized that
some function might be performed by means of its solution. Con-
versely, technical problems may give rise to subsidiary scientific
problems which then become deep and fruitful investigations in their
own right. Indeed, it can be argued that without this sort of 'hybrid
vigour', as well as that resulting from an analogous interaction
between different scientific disciplines, science would become
sterile. But the creation of new scientific problems cannot be man-
aged in an institutional manner as neatly as the exploitation of
results; and so it would be dangerous to the integrity and survival of
science if it were forced to seek its inspiration exclusively from the
problems passed to it by technology. Also, in the long run the rela-
tions of science and technology to the external world will be very
different. The process whereby a new field of science liberates itself
from its technical origins and becomes governed by the advance-
ment of knowledge for its own sake, is entirely legitimate, and
necessary for the continued growth of science. A field of technology
has no such freedom, of course; if it does not continue to contribute
to the performance of functions in the real world, then it has lost its
reason for existence and support.
We have seen that the classes of 'science' and 'technology' are
continuous, and indeed overlapping. The continuum extends through
technology, to more traditional sorts of work: invention and en-
gineering. The defining characteristic of invention is its novelty:
either the creation of a device to perform a new function, or one
which performs an existing function in a new way. In engineering,
the element of novelty is not so strong; here the task is the cre-ettion
330 Science;n the Modem World
of a device to perform a known function in new conditions. Of
course, a significant work of engineering will have a strong element
of novelty, and may give rise to a host of inventions; here, as in
any other sorts of work, the distinction is based on which problem
is considered as fundamental, and which as subsidiary. Generally
speaking, invention relates to the earlier phases of a technical prob-
lem, involving the conception of the device, while engineering is
more concemed with making it work. Whether a particular problem
should be classed as 'technology' rather than as invention or en-
gineering depends on the demarcation adopted; that of Lord
Bowden, that 'an art may become a science if it is concerned with
less than about seven variables', is useful. 9 For it reflects the fact
that in a real problem with many interrelated components, the
abstract categories of scientific argument will not correspond suffi-
ciendy closely to the external world, to enable a technical task con-
ceived in their terms to be accomplished. This is not to deny that
on occasion a technologist may derive useful information from, say,
a multivariate analysis on fifteen ilI-defined variables; but unlike a
scientist in an undemanding field, he does not enjoy the luxury of
of being able to consider such materials as an adequate solution to
his problem.
Because the solution of technical problems ofany of these sorts is
govemed by functions and by criteria of quality assigned externally,
those practitioners who rise above routine work do not have the
same freedom as scientists to develop a specialized competence and
style, ignoring every aspect of a problem that does not suit their
taste. In this particular respect, good work in the solution of tech-
nical problems is more challenging to the intellect and to the imag-
ination than such work in a self-contained field in a matured
scientific discipline. Faced with a particularly refractory bit of the
natural world in his experiments, the scientist can choose to move
on to another problem; while the engineer in the same situation
must fight it out right there, also knowing that the cost of defeat is
far higher than a few months of his personal laboratory time. A
good 'R. and D.' man in industry may find himself engaged in
problems which run right across the spectrum from pure scientific
inquiry to craft-based engineering; and in turning from one to the
nett, he must have flexibility of oudook and diversified skills, to
9 B. V. Bowden, PrOlOsMs ftw 1M DtwIDjmInII of 1M MII1I&Mster Colltft of Samet
1I1IIl1eelntoloo (The College, Manchester, 1956), p. 48.
 Technical Problems 331
apply technique and judgements appropriate to each case. The real
intellectual challenge of good technological work is known to those
who teach technology, and who try to recruit good students to it.
It is then a paradox that English univerity students in technology
are, on average, academically inferior to their colleagues in science;
and so it would seem that technology can make do with inferior
talent. The simple solution to this misleading paradox is that out-
side the traditional engineering tasks, recruitment to technology is
independent of the title of the university degree, and there is then a
sorting by ability, the more able becoming technologists, and the
less able, technicians. I 0

Social Aspects ofTechnical Problem-Solving

The social aspects of work in technology have important differ-
ences from those in science; although the great diversity in the
problems solved, and in the institutional setting, makes neat con-
trasts very difficult. In general, the basic form of property created in
the work is a contribution to the 'art', and as such is registered and
protected formally through a system of patents. But such property
usually belongs to the employer of the man who actually created it,
and for the establishment of his personal intellectual property in a
solved problem, he must use social mechanisms more similar to
those in science. These may be formal, as the publication of a paper
in which he exhibits the solution of the technical problem or a
related scientific problem, or purely informal, through an inter-
personal channel of communication with colleagues. A great variety
ofstrategies will be adopted, which depend not only on the character
of the work done, but also on the regulations and traditions of the
different institutions and occupational groupings. I I
The mechanisms of quality control in technology will generally
be different from those in science, and require a closer analysis. As
we have seen, every solved problem is subjected to tests ofadequacy;
but although in the last resort these depend on craft judgements,
10 N. D. Ellis, 'The Scientific Worker' (ph.D. Thesis, University of Leeds, 1969),
has shown that in his sample of several hundred industtial research personnel there is
no significant difference in work tasks or even in attitudes between those with degrees
in science and in technology. I am indebted to Dr. Ellis for many fruitful discussions on
these and other matters dealt with in this book.
liN. D. Ellis, 'The Scientific Worker', discusses this and reviews the relevant
literature, in the section on 'The Context of Non-Academic Research QS.E.s' Wort
Roles', pp. 123-3°.
332 S';m&t in the Modem World
there is usually a more explicit framework, related to performance,
in which these operate. Similarly, the value of a problem, when it is
conceived purely in terms of its benefit to the person or organization
sponsoring the work, is capable of assessment in largely quantitative
terms. The assessment of the value of a new problem will naturally
be speculative, and sometimes highly theoretical in itself; but it does
not depend on such indefinable judgements as the contribution to
the solution of future problems which do not yet exist.
In general, then, the criteria on which the judgements of quality
are based are simpler in principle in technology than in science. The
function to be performed provides an objective basis for criteria
of adequacy; and the purposes to be achieved, in the way of finan-
cial return, do the same for value. Even those judgements which
in science are so difficult and so sensitive to distortion, as the
choice between investment in totally different fields, are here
rendered straightforward by the common standard of commercial
success. Hence these controlling judgements do not require so much
in the way of accumulated wisdom and statesmanship among
leaders to remain appropriate at least in the short run. Similarly,
the degree of ethical commitment required for the maintenance
of quality, will be less at each level of the structure of direction and
control, than in science. This is not at all to say that scientists are
ethically pure while technologists are impure; rather, that in tech-
nology the various components of the social mechanisms for the
enforcement of quality control are in principle simpler. 1z The
mechanism is then less dependent on ideological and ethical com-
mitments, and can be managed adequately on a basis more similar
to that by which society at large keeps going.
However, the solution of technical problems is not the same thing
as business enterprise in a competitive market; there are important
situations where the simple test of profit is either inappropriate or
irrelevant to the assessment of the quality of a technical problem or
its solution. Even in commercial products, competition by quality
and price is distorted when one firm dominates the market or a
handful share it out between them, selling indistinguishable objects
and competing only in their advertising. Innovations which threaten
such tidy arrangements can be killed at birth by a variety of means,
12 N. D. Ellis, 'The Scientific Worker', shows the important respects in which
industrial Q.s.E.'. do DOt form a 'profession' in any of the accepted senses of the term;
see cit. 6, 'Professionalism in Practice: Its Meaning and Significance'.
 Techn"alProb~s 333
given the superior resources of the established firms to those of any
potential competitor. The suppression of inventions was a more
popular theme before the Second World War, perhaps because the
general depression in business affairs made speculative capital less
available. But the belated emergence of alternatives to the petrol-
engine in the later 1960s, under the political pressures resulting
from recognition of the problem of atmospheric pollution, indicates
that the process ofentrenchment and stagnation of technology is still
with us.
Moreover, there is a large and increasing number of situations in
which there is no effective competitive market to make its contribu-
tion to the testing of adequacy of a device. This will include unique
productions, such as roads, bridges, dams, and large buildings;
once constructed) they occupy their niche for some considerable
time at least, regardless of how well they perform their function.
Similar to these, in principle, are products of which only one model
can be in operation, as is the case with much military equipment.
Those who use such devices may well have their opinions about their
quality; but there is no automatic mechanism whereby the aggregate
of these opinions can be translated into a recognized penalty or
reward. In this way, important parts even of the 'market sector' of
the capitalist economies, suffer from the defects now being recog-
nized in the socialist economies. For the assessments of quality
must be made formally, through institutional channels, and are
thereby subject to certain characteristic weaknesses. First, the
assessments will be derived as conclusions of arguments framed in
scientific categories, in which the evidence is derived from tests.
This work is carried out by experts acting on behalf of the pur-
chaser; and if he is so minded the supplier can always find other
experts to controvert their assessment. In all but extreme cases,
there can be genuine differences of opinion; and when the experts
disagree, there is little hope for the helpless layman, who lacks both
the basis for comparison and the opportunity to switch to another
brand. Also, the institutional framework in which the work of
assessment is done may itself be subject to distorting pressures,
acting through political channels or directly. The most direct
method is through controlling the controllers: the purchasers'
testing facilities are either weakened or subverted by the supplier,
so that he can effectively control the assessments of quality at every
stage of a contract. Such practices are best documented for the case
334 Scimee in the Modem World
ofUnited States Government procurement. There, what is popularly
called 'science' (mainly military and aerospace projects) is treated
like any other commodity being sold to the government, and all the
available profit-maximizing techniques are applied, vigorously and
ruthlessly. 13
In conditions such as these the profits earned by a particular
device are by no means in any direct proportion to its quality;
indeed, in some significant cases, the proportionality seems to be
inverse. 14 These abuses are particularly likely to occur in the most
sophisticated, speculative, and expensive sections of technology, as
military and aerospace, where gigantic sums of government funds
are involved, the competitive market is absent, and political con-
siderations of prestige and 'defence' weigh heavily in the assessment
of projects. In these conditions we can speak of a corruption of
technology; and since it occurs in those sections recognized as lead-
ing, in finance, growth and prestige, its damaging effects can easily
spread. ls Moreover, there is no inherited ideological resistance to
IJ On this, see H. L. Nieberg, I" 1M Nil"" of Stimee.
14 In '''Secret Study" of U.S. arms flops', Tile Cur,""" (27 January 1969), Richard
Scott describes a WtnIIi"fltm Pon story of 26 January about an assessment of military
aircraft missiles in a document then circuJatiDs in Government circles. It was alleged
not only that mstI were double or triple thole estimated, and contracts two or more
yean delayed, but that most of the 'systems' studied were poor in performance: only
four of thirteen worked to 75 per cent of specificatioDs, and four others were discon-
tinued or cancelled outright.
15 One might raise an interesting philosophical question, u to whether shoddy
weapons systems are moraDy superior to effective ones when they are intended for usc
in an unjust war. There is the precedent of the slave-labourers in the Nazi armaments
factories who practised a quiet sabotage by producing substandard work whenever
possible; but it is hard to cast the major American aviation companies in the role of
scc:rct agents for the Vtettong. Of less interest for moral philosophy, but of great im-
portance for society, is the possible spread of corrupted technology into the ordinary
market sector. It is one thing to recognize that mass advertising is generally meaningless
pap; but it is Car more serious ifone cannot have a fair certainty that a device will func-
tion properly at all, or that the manufacturer or merchant will show any interest in
repairing or Iel'Vicing it. It is difIicult to imagine the state of our civilization if 'shoddy
technology' became ubiquitous; I think it would be demonlizing in every way. It is
impossible to say whether the qua1ity control on ordinary manufactured goods is less
effective now than in earlier times; subjective impressions here are extremely unreliable.
There is some evidence that American technology is becoming somewhat tattered at
the fringes. A report from New York, after the minor fiascos with the Boeing 747,
explained the lack ofpublic concern by the fact that in New York at least, no one expects
anything to work; and t;nDIpOrt and communications systems least ofall. See the article
by M. Leapman, Tile Times (28 January 1970). A report ofa decline in standards in a
very sophisticated technology is given by I. Goldman in a letter to Snmee, 167 (1970),
237. Citing several instances of misJabe1led and contaminated radioactive chemicals,
he commented: 'Until m:endy, such slovenliness was confined mOldy to companies that
 TechnkalProbknu 335
the technological equivalents of entrepreneurial and shoddy science.
The traditional engineering fields do have professional institutions
which provide sanctions against dishonest and disastrously incom-
petent work. But these apply only in situations where the engineer
is a true professional; in the large, institutionally ill-defined area of
'R. and D.', where the work is done by technological employees,
there are no effective professional controls. And since commerce
does not provide an effective 'hidden hand' for the maintenance of
quality in these important sectors, the social problems of direction
and quality control can become even more acute in technology
than in science itself. These problems exist independendy of those
of the social and ecological effects of technological innovation, which
I earlier described as 'runaway technology'; although they will
naturally appear together in practice.
One can speak of the quality of the solution of a technical prob-
lem, quite independently of the commercial success of its result.
We do this when rendering honour to great inventors and engineers
of the past; and it is necessary when assessing the quality of any
earlier phase of a full technical problem, as the specification of
operating characteristics, or design. Here, the problem is set by the
existence of a variety of requirements on the solution, which are
independent and perhaps also incompatible. The specification of
the function of the device will impose such constraints on its
operating characteristics; and when these latter are given, there will
be analogous constraints on the design. In one sense, every solution
must involve a 'design compromise'; it is impossible to satisfy all
the components of the solution to the maximum degree. A well-
known example of this, is the array of 'ideal aeroplanes' seen from
the different points of view of parties concerned; the operator will
have seats crammed inside and perhaps tacked on outside too, the
airframes man will have wings big enough for a glider, the main-
tenance man will have all the services running along the exterior,
and so on. 16 And in a routine exercise, one will try to achieve a
enjoyed earned reputations for unreliability. The worrisome aspect of this problem is
that the more reliable suppliers are now allowing this decline of standards.' Con-
firmation of this tendency was supplied by R. W. Fuller in a letter to Slimee, 167 (1970),
1562, 1564- He offered to provide the name of the supplier, and the lot numbers of the
chemical, to anyone in doubt about his own sample. Neither letter mentioned the name
of any offending firm; but the latter came from Lilly Research Labontories, whose
firm is thereby protected from suspicion.
16 I am indebted to Mr. W. Houghton-Evans for this example, as well as for man)'
stimulating discussions on the problems of design in tedmology.
336 Science;" the Modern World
solution which separately satisfies each of the aspects of the prob-
lem, while giving excellence to one or a few. Work of higher quality
is done when the solution ttanseends a mere compromise; and by
thinking the problem through as a whole, the technologist conceives
new relations between the components of a solution, or new means
of satisfying several of them simultaneously.
As in science, work of higher quality is necessary, although in
itself not sufficient, for the solution to a technical problem to be
more than ephemeral. Although the criteria of quality will generally
be different from those of science, there are processes of selection
and ttansformation of results closely analogous to those of the social
phase of the development of scientific knowledge. We have already
discussed the patterns of evolution of some devices, in connection
with tools in science. The analogue to the stability and invariance of
a scientific result is the successful performance of its assigned func-
tion, and the adaptibility to new functions, of a device. What can
eventually survive through all changes of context and function, is
the principle of a device, which becomes a basic, elementary piece
of information, used in a variety of forms in the solutions of tech-
nical problems at all levels and all degrees ofscientific sophistication.
In this way, technical problems as well as scientific problemscan yield
a permanent contribution to our knowledge of the world around us.
From this analysis we can see why a 'classic' design, in which
form and function are perfectly united, is as unique and unrepro-
ducible as a classic in aesthetic creation. For as the inevitable
changes in function (including aspects of performance) demand
change in design, the original unified conception must itself be
modified. This may be done creatively, so that several 'marks' of
a device maintain or even enhance its original excellence; but it is
more common for changes to be made piecemeal, so that the later
'marks' become an incoherent pastiche of ad hoc solutions to new
problems; eventually only the name and the memory of 'the way
they used to build them' survive. The original conception is then
essentially obsolete; and the craft of design (and with it the per-
formance of the assigned functions) will decay until a new creative
synthesis is achieved. This phenomenon can be seen in automobiles,
and also in later 'stretched' versions of aeroplanes. I?
17 I am indebted to E. Paul Adler for this important point, which shows that techno-
logical propea is, even on the small scale, DO more smooth and inevitable than scientific
progress.
 Technical Problems 337
Varieties ofTechnical Problems
Although this discussion has concentrated on technical problems
concerning physical objects and processes, the same analysis holds
for a wider class of problems. For example, the largest and most
important technical projects of the present include some concerned
with 'information'; and in computers, the 'hardware' and the 'soft-
ware' are correlative parts of the complete system. Thus, the per-
formance of a function need not require a physical device; the
organization of information, or of human activities, can constitute
the solution of a genuine technical problem. Such solutions can vary
in quality from the merely competent to the very deep. For instance,
the systems of law and administration developed in the Roman
Empire are a permanent contribution to human knowledge, as much
as any principle of a physical device. There will, of course, be
characteristic differences between technical problems of this ex-
tended sort, and those involving physical things; in particular, when
the objects of inquiry and manipulation are themselves possessed
of consciousness and purposes, they must be handled differendy
from inanimate things. But such differences exist within science
itself, especially if (as in other European cultures) the concept ex-
tends across all fields of scholarship. The 'soft' technologies will
have certain characteristic differences from the 'hard' ones; their
associated sciences will usually be less matured, and in any event
incapable of providing a doctrine which completely specifies all
routine problems. But these differences, too, are of degree rather
than of kind.
Thus, the distinction between technical and practical problems
is not between non-human and human objects of inquiry and
manipulation; but between the determining final causes of the task.
In technical problems, the ultimate purposes to be served are
reckoned with only as general constraints on the solution, com-
monly through the question of whether it will yield a profit; while
in practical problems these purposes determine the functions to be
performed, and from them the individual tasks to be accomplished.
For example, the work of a surgeon is a practical problem; for unless
he restores the health of the patient, his operation, however, brilliant,
will have been a failure. Conversely, the work of administration, at
any but the highest level, is of a technical character; for although
people are involved as objects, the immediate problems are deter-
mined by the particular functions to be performed, while the
338 Science;" the Modern World
ultimate purposes, even if they condition the work, are fixed and
remote.
The correct identification of the distinction between technical
and practical problems is of great importance for understanding the
human aspects of modern physical science and technology. If one's
language, in the term 'science', forces one to assume that work with
physical things has an autonomous system of final causes, and that by
conttast work with people is automatically concerned with welfare,
then there is no framework in which these problems can be analysed,
much less solved.
 13
PRACTICAL PROBLEMS

WE now come to that class of problems, conceived in terms of in-

tellectually constructed objects, where the ultimate purpose of the
task directly determines the goal. We may define a practical problem
as a statement of a purpose to be achieved, whose means are to be
established as the conclusion of an argument, with a plan for its
accomplishment. In some respects, such practical problems are the
most difficult to solve, as well as including some of the most im-
portant tasks facing our society, those coming under the general
category of 'welfare'. It is to these that the 'social sciences' are
directed, as well as the embryonic sciences of 'the environment'.
There is a class of practical problems which were the subject matter
ofthe first learned disciplines to achieve maturity and social effective-
ness: those which are the concern of law and theology. In the prac-
tice of the law for example, the particular tangle in which a client
finds himself is immediately translated into the categories of legal
theory; once an interesting case is underway he is reduced to the
status of a name-tag; and yet the goal of the advocate's task is not
to establish novel legal doctrine, but to win the case on behalf of
his client. However, our main concern here will be with those large-
scale practical problems whose solution involves the execution of
practical projects, analogous to the technical projects necessary for
the complete solution of a technical problem; for it is these that
require understanding most urgendy if their many pitfalls are to be
avoided.
The Cycle ofInvestigation, and CluJrflcteristic Pitffllls
As in the case of the other sorts of problems, a practical problem
has a phase of gestation, when it is a problem-situation: an aware-
ness that things are not as they should be, but no clear conception
of how they might be put right. Once it comes into being, its cycle
340 Science in the Modern World
of investigation may be described by five distinct phases: definition;
information and argument; conclusion and decision; execution;
and contro!.1 Such a division of phases may be used to describe the
accomplishment of any task; once the goal is set, one proceeds to
consider the possible ways of fulfilling it; then decides between
alternatives; proceeds to the operation; and during the actual work,
supervises it to ensure that the goal is being satisfactorily fulfilled.
The phases are quite sharply differentiated in practical problems.
Thus the purpose to be achieved is necessarily stated in intellectually
constructed categories, and so it is necessary to have the results from
an inquiry of a scientific sort on which to decide the best means for
its achievement. Also, the phase of execution of the work will
usually be distinct from that of decision, in it~ ~gents as well as in
its techniques; and finally, where there is ani division of labour,
the task ofcontrol cannot be neglected. Indeed, control may become
so elaborated that it constitutes a technical problem, in the strict
sense, with its own social mechanisms for each phase of its cycle,
including that of controlling controllers in a variety of ways.
Comparing this cycle to that of scientific and technical problems,
we notice certain systematic differences. Ultimate purposes) which
are remote and diffuse in science, become a part of the criteria for
the controlling judgements in technology, and here they determine
the goal itself. Outside those limited fields where purposes are
capable of being handled in terms of accepted explicit intellectual
objects, the framing of a new practical problem is an essentially
creative act. Whereas the setting of a problem in science involves
the partial and tentative specification of a conclusion about artificial
objects in a self-contained universe, and a technical problem in-
volves imagining a device to perform a preassigned function, here
the specification is of a state of affairs in human society which does
not yet exist. Each of the controlling judgements of feasibility, cost,
and value involves a multiplicity of factors, few of them reducible
to quantitative or routine assessment. Moreover, the statement of
the goal of the problem, as well as those of the controlling judge-
ments, presupposes a social and moral philosophy. This may be
I This cycle might be compared with that given by J. Bny, Decision i" Govmnnent
(GoUancz, London, 1970), p. 269: 'objectiTe - model- hypothesise control action - pre-
diet - vary hypothetical control action - revise prediction - repeat to select optimum -
act - observe behaviour - refine objective - revise model - update prediction and
oprimigtion -act again - continue to observe - all in the light of changing external cir-
cumstances'•
 Practical Problems
341
implicit and informal, and may seem obvious common sense to its
proponents. But it is an ideology, a universe of reality and value,
which itself is incapable of simple testing and scientific control. It
will usually be articulated in a folk-science, whose peculiar features
we will discuss later. The work of investigating and solving the
practical problem, will necessarily be done within the framework of
that ideology in whose objects the problem is conceived and first
assessed. The categories in which the information is produced, and
in which the argument about alternative means is conducted, will
be determined by the ideology. The decision to be reached, depend-
ing as it does on the conclusion of that argument, is thereby limited
in its range of possibilities. Indeed, once the problem has been
framed in a certain way, the decision on its means of accomplish-
ment may well be completely determined. In this way, the very first
phases ofthe investigation ofthe problem have a closely determining
influence over all those that follow. In technical problems, the rela-
tion is that of a sequence of final causes, each one delimiting the
range of possibilities for the next phase: but here each phase also
provides all the materials for the next, and so delimits it even more
narrowly.
Although the categories in which the practical problem is solved
are so narrowly restricted by its initial conception, this does not
necessarily have the effect that the ultimate purpose to be served is
dominant in all the tasks involved in the phase of execution. Several
features of this work can cause a displacement of the effective goals
of individual tasks into other hierarchies of final causes. The most
universal cause is that all the agents will have private purposes oftheir
own, which are not automatically in harmony with the goals
determined by the ultimate purposes of the work; and the prob-
lems we discussed in connection with social behaviour in science,
especially those of quality control and morale, have their analogue
here. Z Another cause, more noticeable in large practical projects,
is that the work of execution gives rise to challenging technical
2 Studies in the sociology of bureauaacies have analysed these tendencies to 'goal-
displacement'. A very vivid account of a case-6tudy of this process is given by Harry
Cohen, The Dmumics of BWIQualUY (Ames, Iowa, 1965), written in some ways as a
sequel to the basic work by P. M. Blau, Tile Dyumics of BwellU&rMY (University of
QUcago Press, 1963). Cohen's analysis shows the pressures causing a deviation from the
fonnal goals ofthe bureaucracy, and stresses the importance ofthe informal adjustments
made by individuals to their tasks for their self-protection. He offers suggestions for
munteneting these tendencies; but at the time of writing had not mnsidered the use-
fulness of bureaucrats being answerable to their clients.
12-S.K.S.P.
342 Science in the Modem World
problems or projects; the performance of an assigned function
becomes the focus of attention, while the purpose being served
through that function is neglected. This is an entirely natural
tendency; it is easier to specify a function than a purpose, just as it
is easier to specify a goal than a function. Hence it is more con-
venient, and safer, for agents at any level and those who control
their work, to restrict their consideration and responsibility to the
most immediate final causes involved in their tasks. This tendency
for immediate technical problems to displace the initiating practical
problem in the execution of a project becomes very marked as soon
as a stable organization has been formed, and 'welfare invention'
has given way to 'welfare engineering'. This has its own effects, on
the categories in which tasks are defined, and adequacy assessed.
Like the scientist's conceptual objects, these cannot exist inde-
pendently of their use. Although they may have been first conceived
in a philosophical reflection on a social problem, when they come
to be applied to the complex tasks of the project they will need
additional specification for their new functions. This will be in the
direction of appropriateness for the definition and control of the
many routine tasks to be accomplished as conditioned by the indi-
vidual and collective purposes of those involved at each level of a
hierarchy. In time the original purposes which these categories
expressed may well become inexpressible, and hence non-existent,
in the framework of their elaborated, technical meanings.
The operation of all these tendencies, in a large and complex
organization, can give rise to the characteristic features of a 'bureau-
cracy' (in the pejorative sense), in which the original defining pur-
poses are forgotten and the agents concentrate on an internal
political game, treating their nominal clients as irrelevant nuisances.
Thus, although one can establish a clear distinction in principle
between practical and technical problems, in terms of the import-
ance of ultimate purposes in defining the goals of the task, in
practice it is more difficult. There is no simple criterion in style of
working to distinguish between those established organizations
which were initially created to solve a practical problem and those
whose only ultimate purpose was and is the personal benefit of their
promoters and owners. This depressing conclusion accords with the
experience of Socialist societies; and it raises deep problems for
Socialist theory, whose tradition was generally negligent of the
problems of bureaucracy. It would appear that the solution of such
 Practical Problems 343
problems lies either in the reduction of the size of all political and
economic units to the point where bureaucracies are unnecessary,
or for the improvement of methods of accountability of large
organizations to those who exercise ultimate control, and of answer-
ability to those whom they should serve. 3
The two features of ideological origin and bureaucratic execution,
so different and indeed contradictory, combine to produce an
extreme tendency to rigidity in the solution of a large practical
problem, so that (unlike in science or technology) there is a mini-
mum of re-cycling through the phases of the problem. In the earlier
phases, any challenge to the conception of the problem is likely to
meet with intense resistance from those whose personal ideologies,
as well as careers, are involved in its planning and accomplishment.
And once the routines have been set and organizations established,
the whole enterprise has an inertia which is very difficult to over-
come; for the costs, both political and administrative, of any major
changes, "ill be seen by those involved as prohibitive.
This rigidity creates particular hazards. The pitfalls that are
encountered whenever a system of intellectual objects makes con-
tact with external reality, are reduced in incidence only by a tenta-
tive and exploratory approach to that contact, so that clues to their
presence can be detected. Here, the close determination of the
problem by its initial conception removes that warning system; and
so practical problems are particularly prone to great and disastrous
pitfalls. Similarly, the reduced re-cycling leads to the long-delayed
recognition of pitfalls when they have been encountered, and a
correspondingly greater difficulty in repairing their effects. 4 Hence
3 This latter seems to be the mntent of the demand for 'Socialism with a human
face', raised in Czechoslovakia in 1968; and it is also part of the movement for reform
within the Roman Catholic Church. The concept of 'answerability' to those inferior in
an hierarchy, was developed by myselfin a paper, 'Power, Responsibility, Answerability',
delivered to a teach-in on 'The Nature of the University' orpnized by the Leeds
University Union on 12 December 1968•
.. All these hazards arc intensified by the natural tendency of men of vision to devote
their years of power to the solution of the great problems of their youth, despite the
gap of decades during which the problem may well have been completely transformed.
For an example ofthis tendency, applied as an explanation ofCromwell's policies during
the Protectonte, see H. Trevor-Roper, 'Three Foreigners', in Relititm, tile ReftJnlUllion
.1111 Sod.l CIuI",e (MacMillan, London, 1967), pp. 23~3. After describing the
philosophy of the disastrous decade of the 16201, and the changes wrought by inter-
vening times, he mJDJDeDts: 'But Cromwell muld not change his mind. It had been
moulded, fixed and perhaps slighdy cracked, in the grim and lurid furnace of the put.
So now, as Lord Protector, he adopted a foreign policy that was twenty years out of
date: ••• Protestant reunion in Europe, Elizabethan war in the West Indies, and a
344 Science in the Modem World
a common cycle for a new practical project is to proceed from a
grand ideal conception, into a morass of difficulties of execution,
and then either to lurch forwards into a final collapse or to survive,
in a battered state, accomplishing tasks which have little relevance
to the original purpose. or perhaps to any at all. This latter fate is a
likely one even for projects that are well executed; for the time-lag
between initial conception and large-scale execution may be so
great that the real practical problem will have altered radically in
the interval. And the worst feature of such a sequence is that those
responsible for it are usually incapable of learning any lessons from
the experience. The whole process is too complex to provide decisive
tests of particular points of the ideology and folk-science in whose
terms the project was first conceived, and any bureaucracy has
built-in barriers to the recognition of failure in the performance of
its functions. 5
The work of solving practical problems also suffers from the
inevitable absence of a consensus on the criteria by which quality or
success is judged. In technical problems, these criteria are supplied,
to some extent at least, by the objective tests of performance of
function, and of commercial success; and in science it has been
possible, in many important cases, for the members of the small
community associated with a field to develop appropriate criteria
for their work. But practical problems affect a variety of people in
a variety of ways; there will be those who decide, those who execute
the decision, those who are supposed to derive benefits, and those
who incur costs. Such diverse groups will not only view the problem,
top-dressing ofideological mysticism which included the reception of the Jews' (p. 282).
The essay is a magnificent study of the fortunes of Utopian sd1cmcs u they are tossed
in the storms of politics.
5 A particularly gloomy view of this process wu taken by the social theorist Roberto
Michels in his Politi&tJI P"";es: II Soeiolo,i&tJI Slruly of Oli,.,hi'QI TeruJmdes of
MtNlmt Det1uJerM1 (German original, 1911; tnDsIated E. and C. Paul, The Free Press,
New York, 1949). He desaibes an 'iron law of oliprchy' whereby with the advent of a
mass organization, bureaucracy kills the democracy within the movement. With the
Dineteenth-century Socialist movement in mind, he concluded his book with: 'When
democracies have pined a certain state of development, they undergo a gndual trans-
formation, adopting the aristocratic spirit, and in many cases the aristocratic forms,
apinst which at the oullet they struggled 10 fiercely. Now new accusers arise to de-
nounce the traiton; after an en of glorious combats and inglorious power, they end by
fusing with the old dominant claD; whereupon once more they are in their turn attacked
by fresh opponents who appeal to the D&IDC of democracy. It is probable that this auel
pme will continue without end' (p. 408). See R. Nisbett, Tile Sotioloti'QI TrtUlititm
(Buic Boob, New York, 1966; Heinemann, London, 1967), pp. 148-50 for a discussion
ofMicbcls.
 Practical Problems 345
and its effects, in ways depending on what they consider their self-
interests; they will also have different universes of reality and value,
with their corresponding folk-sciences, in which the words used for
the description of the problem relate to radically different things.
In the debates over particular practical projects, the various sides
will inevitably be talking at cross purposes about different things. 6
Appeals to 'the 'facts' are of little use; for 'facts' are even more rare
and evanescent in collections of such informal systems than they
are in science. The situation is analogous to that in an immature
science, but even worse in that the most fundamental objects of
argument, values and purposes, are inherently incapable of being
the subjects of universally accepted demonstrated conclusions. 7
Yet with all these inherent weaknesses, it is necessary to solve
practical problems as I have defined them, if certain sorts of
6 In a study of the historical background to the Q!wry Hill Flats in Leeds, Dr. A.
Ravetz has identified the following series of practical problems associated with housing
and the poor: 'insanitary areas' from the 18308; 'bad housing conditions for the working
classes' by the end of the century; c. housing shortage' at the end of the First War;
cslums requiring clearance' in the 19305; Chousing IS a social service' around the same
time; and covercrowding' in the mid-193OS. The earliest problem received attention IS
a menace to general health conditions, and the relevant function to be performed was to
provide water, sewage, and fresh air. For this last, houses would be destroyed, the fate
of the inhabitants being ofseamdary concern or none at all. Mer the Fint World War,
the first Council estates were built under the slogan of'homes for heroes' and their being
too expensive for the slUllHiwelling poor was an irrelevance. In spite of this, some lead-
ing Socialists interpreted the measure as a step towards 'housing as a social service' and,
on these grounds, bitterly opposed the later policy of giving priority to re-housing the
desperate slum-dweUers. In Leeds, the Labour movement was itself split on this
disagreement about the nature of 'the housing problem' and the split was one of the
contributory factors in the defeat of the Revd. CharlesJenkinson in the local e1ec:tiODS of
1935 after the brief period of two heroic years in power.
7 The common feeling of inferiority of cloistered academics with respect to successful
practical men of a1faits may be assuaged by the experience of Descartes when he left
the world of books and entered that of real action. As he te1Is it, he went out into the
world seeking the way to Truth, 'For it seemed to me that I should find more truth in
the reasonings which a man makes with regard to matters which touch him closely, of
which the outcome must be to his detriment, if his judgement has been at fault, than in
the reasonings of a man of learning in his study, whose speculations remain without
effect, and are of no further consequence to him than that he may derive all the more
vanity from them the further removed they are from good sense, because of the greater
skill and ingenuity he has to employ to make them plausible.' Unfortunately, 'It is true
that so long IS I did nothing but reflect upon the behaviour of other men, I found no
grounds in it for assurance; indeed I perceived in it IS many divergences IS I had form-
erly found in the opinions of philosophers.' Dis&ofll'se on Method, translated by A.
Wollaston (penguin Books, 1960), last page of 1St Part, p. 43. Descartes did draw lessons
from this experience of the real world, ones which could have been (and perhaps were)
derived from a reading of Montaigne; and he could then have recourse to the c1aI#re
""tfll'elJe' of his mind, to achieve his distinctive philosophy.
346 Science in the Modem World
purposes are to be achieved. It is naive in the extreme to believe that
the common sense and good will of all right-thinking people could
suffice for the accomplishment of important social tasks. For when
there are situations where there is a consensus on ends and spon-
taneous co-operation on means, then the practical problem does not
arise in the first place, and the social task is accomplished without
attracting any notice. But where these favourable circumstances do
not exist, any co-operative attempt to fulfil collective goals while
also serving the personal purposes of the agents without any
definition of the tasks can yield nothing but a shambles. Nor can
one rely on the presence of a 'hidden hand' which will ensure that
an aggregate of purely private purposes will somehow be equivalent
to a beneficial public goal. An equal naivete at the other extreme is
found in the beliefthat 'scientific method' can be applied in a simple
and straightforward fashion to this class of problems. 8 This claim
was part of the propaganda for science in earlier times; and the
history of social reform is littered with the failed schemes of socially
conscious intellectuals and tidy-minded reformers who blundered
into the first pitfalls in their way. To explain such failures exclusively
by the stupidity and selfishness of politicians and of other groups of
people is merely to ignore the difficulties inherent in practical
problems and to be just as incompetent on the next attempt.
'Scientific' Aspects ofP'II&tical P,oblems
The sphere of activity comprised in the solution of practical
problems is so vast and complex, that its proper study requires, and
has already created, an extensive research effort. Here we will
restrict ourselves to considering those aspects of practical problems
which relate closely to science. The relation is twofold: the influence
of the specifically 'scientific' character of practical problems as
distinct from tasks; and 'practical' aspects of technical problems.
One may say that a set of tasks constitutes a field of practical
problems, when it is recognized that mastery of some demonstrative
discipline is necessary for their accomplishment. Medicine and law
• Karl Pearson, in Tile c,.... of samet, strongly advocates the application of the
methods and the results of science to social problems. His chosen example wu the
problem oCthe poor; and he used Waissman's theory oCthe germ-plasm (with reserva-
tions in his account of the theory but not in his pnetical conclusions) to condemn that
iIl-conceived philanthropy which encourages such 'inferior stocks' either to increase their
Dumbers or to mix with tho8e IUperior to than. (EftI'JIII8D eel., introductory, paragraph
9t pp. 27-30-)
 Practical Pr(}blems 347
have been so recognized for many centuries, and they have provided
the model for the 'learned professions'. This recognition is not a
simple thing; the group of practitioners must convince both the
State and the public that a particular set of purposes can be achieved
only by their efforts, and that competence for attempting these
tasks must be assessed and certified by themselves. The problems
of the would-be professions in the modem world were anticipated
by medicine over the course of centuries; since the training for
'physic' was largely vacuous in content, and practical competence
was not tested to any degree of rigour, the physicians had to contend
on all fronts with a variety of competitors. 9 Indeed, the learned pro-
fessions did not have a very strong claim to their monopoly of
practice until quite recendy.
It was during the last century that the welfare of the population
came to be seen as involving practical problems as I have defined
them; and with the movement for the creation of corps of skilled
agents of various sorts, and their proper training and certification,
the professions themselves, old and new, gradually put their houses
in order. Many of the projects that were established required the
work of new sorts of technicians, in particular those involved in the
detailed work of inspection and regulation. Although their tasks are
sometimes defined in principle by certain ultimate purposes (such
as the protection of some or all of the population from bad food or
bad working conditions), their working contact is with physical
things (or symbols) rather than with individuals. But there is another
class of skilled agents, who (in principle) serve individuals to im-
prove their welfare; in this respect they have an affinity with pro-
fessionals, and we can describe them as 'practitioners'.
Both technicians and practitioners are restricted in their work to
routine tasks; difficult cases are referred to superiors, and genuine
innovation is reserved to the highest level in a bureaucratic project.
Their training will be correspondingly straightforward and detailed,
and less theoretical and 'liberal' than that of those destined for
higher positions. Since they are put into contact with a reality
9 See R. H. Shryock, The Development of Motkm Medicine, for a general history of
this problem. The sbUggles between the CoUege of Physicians and its rivals (apothe-
caries and 'chemical physicians') in the seventeenth century is particularly well docu-
mented; see Sir George Qark, History ofthe ROYillColltte ofP"ysicilltU, vol. i (Oarendon
Press, Oxford, 1964); and C. Webster, 'English Medical Reformers of the Puritan
R.evolution: a Background to the Society of Clymical PhysitiaDs', ~, 14 (1967),
16-41•
348 Science;n the Modern World
external to that of the artificial categories of the project, they are
exposed to the divergence between theory and the external world.
This can happen even to technicians, who may find themselves
deprived ofthe resources with which to do a proper job of inspection
or regulation on behalf of the public. 10 Practitioners may find them-
selves in an even more conttadictory position; for the administtative
case-law which they are required to enforce may be designed more
for convenience and parsimony than for the welfare of their clients.
By the nature of their work, they are accountable to their superiors,
but they are liable to be answerable to their clients; and the two
sorts of goals imposed on their tasks may be in direct opposition.
However, we shall see that their ttaining for their work can provide
a protection against such discomforts. There is a very important
class of skilled agents involved in practical problems, who fall
between technicians and practitioners on the one hand, and true
professionals on the other. The 'experts' are not restricted to routine
tasks, but will study genuine problems, involving the exercise of
judgement and the deriving of a conclusion from an argument. If
they work directly for a client, they may merely suggest a decision
and leave the execution to him; or they may undertake that work
as well, depending on conditions. The objects of their work may be
human, non-human, or a mixture. The expert is distinguished from
the true professional by his position, entailing the controls exercised
on his work. For he is an employee of a firm, accountable to his
superiors there who have the powers of penalty and reward over
him. His answerability to his clients is very slight, and to his col-
leagues in his specialism, not much greater. This difference is of
course crucial; and is recognized in our language. Given two men,
of identical skills, and ostensibly offering the same service, the
employee is an 'expert', and the independent agent, the true pro-
fessional, is a 'consultant'.11
10 An example ofthis phenomenon is provided by A. D. Woolf, 'Industrial Accidents:
Can we Stop this Suffering?', letter to Tile Times (Business News Section), 24 March,
1968, p. 28. He quotes Mr Plumbe, the Chief Inspector of Faetones: 'The Inspectorate
bas never aimed at, and certainly hu never achieved, • rigorous enforcement of the
(Factories) Act such u a Teutonic country might attempt.' He comments, 'The pathetic
Dumber of prosecutions set against the annual toU of accidents amply confirms his
statement.'
I I The difference between 'expert' and 'consultant' becomes quite clear from examples

of situations where they work in competition. Thus: 'Recendy, in California, a small

poop of independent entomologists hu found that advising cotton growers can be a
profitable business. Growers contract to have their fields checked once a week by these
 P'II&tual Problems 349
As we have seen, an important class of technical problems are
incapable of being subjected to quality control through the auto-
matic mechanisms of the market; and when disputes arise because
some person or group considers their welfare to be neglected or
injured, experts are called in by both sides. In some ways their role
is analogous to that of an advocate; but they do not operate within
the sophisticated etiquette and ethic of that profession whereby a
man can argue for his client without losing his integrity. Rather, the
expert tries to argue (perhaps with sincerity) as if he were a scientist,
establishing his conclusions on supposedly known and irrefutable
facts. The absolute loyalty ofexperts to their organization can some-
times be quite touching; even when a project is revealed to have
stumbled into the most ghastly pitfalls in technical execution, the
firm can usually find an expert to assure the public that the matter
is really in competent hands. As a result of this aspect of the experts'
task, their assertions come in that class to which is applied the maxim,
'Never believe anything until it's been officially denied.'12 It is
perhaps less serious that the general public is ignorant of these
limitations on the reliability of expert pronouncements, than that
their superiors in positions of decision-making must, by bureau-
cratic etiquette, place some reliance on their conclusions and
recommendations. 13
EveR in the case of technicians and practitioners, the tasks of
men for a per acre fee that is usually less than the cost ofone insecticide treatment. They
furnish the same service for a fee as the industry salesmen furnish free. The difference
is that they are selling only the service and not the chemicals.' Kevin P. Shea, 'Cotton
and Chemicals', Srimtist tiM Citiun, 10 (1g68), 209-20. (This journal was subsequendy
~ded EtrDirtnmle1ll.)
la A classic case of 'official denial' occurred after one comer of a point-block of
London couneil flats (Ronan Point) had gone down like dominoes following a gas
explosion in one upper room. See TIte Times, 17 May 1968.
13 'There are three roads to ruin.: women, gambling and technicians. The most
pleasant is with women, the quickest is with gambling, the surest is with technicians.'
M. Georges Pompidou (quoted in The SUtUliIy Telegraph, 26 May 1968). Thus the
Pacific Gu and Electric Company fought a 10Dg batde apinst conservationists and
geologists to establish a nuclear power station at Bodega Head, north of San Francisco.
Only after five years ofargument was a proper seismic survey done on the site on which
a large hole had already been bored, and this showed the bedrock to be disastrously
faulted. It would appear that the directors of this utility, so shrewd in other respects,
must have believed what their 'experts' were telling the public about the suitability of
the site. See Joel W. Hedgpath, 'Bodega Head_ Partisan View', Bulleti" oftlte Atomi,
Seimtists, 21 (March 1965) 2~, and Sheldon Novick, TIte Careless Atom (Houghton
MiftIin, New York, 1969). Examples of the way that 'experts' discIw'ge their duties, in
England no less than in America, are not hard to come by; but under present law one
hesitates to mention these in print.
350 Science in the Modem World
training and certification require a formal body ofdoctrine presented
as a 'science'. Without this, the work would be recognized as merely
a craft; and with skills transmitted purely interpersonally, by ex-
ample and precept, it would be impossible to enforce any uniformity
of standards for certification, or of practice on the job. A stable
bureaucracy, with its complex internal technical problems of
organization and control, could not operate under such conditions.
Hence if an appropriate science does not exist for a particular
practical problem, it is necessary to invent one. Once it is available,
its further development can perform several valuable functions for
its field. On the social side, the claim to possess a body of theoretical
docttine, necessary for competent practice, helps the occupational
grouping of practitioners and experts to consider itself as equivalent
to a profession; by this technique, it can demand status for its
present members, and tty to influence the flow of certified recruits
in the direction of smaller numbers and superior social origins. 14
Also, the theoretical discussion of the body of docttine, and the
activity of research in its terms, enhances the claim to truly pro-
fessional status, and also provides the basis for a demand that the
subject be recognized as properly belonging to the world of higher
education. Although the body of doctrine in question may be nearly
vacuous, an incompetent rationalization of a craft practice, it can
still perform several functions in the internal workings of the field
of practice. As it becomes elaborated, it can become the dominant
or sole framework for the definition of the tasks accomplished by
practitioners and experts, excluding common sense altogether from
14 For the penetrating amunents of Max Weber on the role ofeducation in the main-
tenance of a burcauaatic Bite, see R. Bendix, Mu Weber, tm I"telleettUll PortrtJ;t
(Doubleday Anchor, New York, 19M), pp. 229-3°,461. Robert M. Hutchins tells the
following story: 'I shaD nefti' fcqet-though I have often med-the last meeting of
the Big Ten [Midwestern Universities] presidents I attended, fifteen years ago. The
President of the University of Michigan said "Say, I want to ask you fellows, what are
we going to do about embalming?" He went on to report that the embalmers in his
state wanted to become a profession for the double purpose of limiting competition and
nising their social standing. This they proposed to accomplish by establishing a school
of embalming at the University of Michigan and requiring all practitioners to have a
degree from the school before being permitted to embalm any resident of the state. He
was consoled by the president of the University of Minnesota, who assured him that
"the embalming program at that great university had not interfered with its smooth
operation".' See Oerk Kerr and othen, The U,,;wrsit] ;" AmtrietJ (Center for the
Study of Democratic Institutions, Santa Barbara, California, 196'7), p. S. The Michigan
embalmers were aiming for truly professicmal status by requiring their recruits to train
at the prestigious arts and scienca University of Michigan nther than at the land-
grants Michigan State University.
 Practical Problems 3S I
their conception of their work and from the criteria for its quality.
This is convenient both for the agents and for those who exercise
decision and control over them; the raw world of untidy and con-
ttadietory values and purposes of real people can be henceforth
ignored in favour of the neat conceptual 'world of the field. In this
way, practitioners at every level are at minimum risk; and the ulti-
mate purposes of the work, to the extent that they are reckoned,
are so defined in the terms of the theoretical structure that they are
automatically served every time a unit task is accomplished. ls
The situation of the true professional presents certain barriers to
this invasion of his field by science, or pseudo-science, which are
conveniently absent in the c3se of the bureaucratic practitioner and
expert. For the professional is approached by a client, who can
choose the best means of having his purposes achieved. If the client
goes away dissatisfied, he can tty a different man the next time; and
the profession as a whole always faces competition on the fringes of
its field, from unqualified practitioners who advertise themselves as
achieving the same purposes by different means. But the bureau-
cratic practitioner almost always enjoys a monopoly; ifhe distributes
welfare on behalf of the State, his clients, through ignorance or
poverty, cannot turn elsewhere; if the task is regulatory in any way,
his monopoly will be enforced by the State; and if his service is
provided by a private firm, it is likely that it will be one of a few
which together dominate the market, and have identical styles of
operation. Hence the practitioner need know nothing at all, from
15 Tendencies in this direction are described by Barbara Wootton, SoeW Stietltt 11M
Sotilll Pllt],oloD (Allen & Unwin, London, 1959). 'In some cases this emphasis on the
hidden psychological issues that are supposed to be uncovered by what bas come to be
mown a the "casework process" bas gone so far IS to lead to almost deliberate dis-
regard of the practical problems which were the immediate occasion of the relationship
being established' (p. 278).
Another very important function of the theoretical doctrine in whose terms practi-
tioners conceive their problems is as a folk-6Cience for themselves, enabling them to
cope with the intolerable situations they encounter. These may arise in interactions with
people whose values and way of life are abhorrent and incomprehensible, or whose
suffering is so severe that the practitioner cannot identify with it. The natural defence is
to dehumanize such 'patients', in the former case by applying moral judBanents to their
situation; such indeed was the attitude ofmuch 'charity' in former times. The theoretical
doctrine offers a means to a more sophisticated and depersonalized defence, for in its
terms the moral and emotional threat is explained away. The practitioner can then allow
his natural sympathies to operate without danger to his own person. Needless to say,
the formation and adoption of a doctrine is governed by many other functions as well,
and so it is not always well suited for this one. I am indebted to Dr. H. W. S. Francis for
this important point.
352 Science in the Modem WorltJ
direct human contact, about the purposes ofthe people he is claimed
to be serving. To him they can be merely the sum of the theoretical
attributes which he memorized in the course of his ttaining; and the
assessment of his fulfilment of his goals will be based on these cate-
gories, as modified by an administrative case-law.
The new technology of information has recently provided tools
which supplement the work of the bureaucratic practitioner, and in
some cases even make him redundant. The routinization of tasks is
pushed to its logical extreme, when every particular case is com-
pletely characterized by an array of holes on a punch-card. This
technique not only saves manpower and money; it also provides
materials convenient for analysis by that most fashionable of
sciences, mathematics. With such data, research can become a
routine, and the members of the 'scientific' field associated with an
area of practical problems can generate the pseudo-property of
research tides with no more labour than the hiring ofa programmer.
The human beings whose purposes are involved in the field, are
now at one remove further from the perceptions of those who con-
trol their welfare. Even if their particular situation is capable of
translation into the categories in which the practitioner or expert
operates, they are at risk of being overlooked by the systems analyst
who constructs the computer program, or even of being the
victim of an error by the key-punch operator. Should that occur,
they cease to exist as anything remotely resembling their true
selves, and either vanish as unclassifiable or re-appear in foreign
garb. There are few defences available to the individual against such
a destruction of his identity; as a mute protest against the machine,
one can fold, mutilate, find spindle one's punch-card if one is
fortunate enough to handle it;16 but otherwise one is truly up
against a System.
The natural propensity of bureaucratic practitioners and experts
to reject and condemn any particular case which falls outside their
routine categories is exaggerated and freed from constraints when
the 'case' is not even a file of papers but only a punch-card. The
computer is, after all, only a very rapid clerk which strictly 'works
to rule', and which makes many of the rules itself by its physical
limitations. More sophisticated categories of its objects require an
16I am indebted to Dr. Robert S. Ravetz, of Philadelphia, for this technique. It
to bring one', grievance to the attention of those who can deal with it, unlike a
IeI'ftI
mere letter wbicb is automatically shunted to a 'complaints' bureaucracy.
 Practical Problems 353
elaboration of procedures which is costly to achieve, and whose
operation will waste the computer's precious time; and non-standard
cases, spat out by the computer, are a serious nuisance. Hence the
natural tendency is to replace any judgements of attributes by
simple sets of parameters, and in case of doubt to apply 'fail-safe'
to the benefit of the organization. It is for this reason that the
proposed 'data banks' are a threat to civil liberties. Their materials
will be raw data from a variety of disparate sources; it will be con-
verted into information by automatic procedures; and then inter-
preted as evidence for practical conclusions in the context of a
computerized bureaucracy. Any citizen will then be at risk of being
victimized, with no possibility of redress, because of some deviance
(real or erroneously imputed) from some expert's imagined 'norm'. 17
It is always important to remember that such developments as
these arise from natural causes and also bring great benefits along
with their dangers. Practical projects come to be when large-scale
purposes cannot be served in any other way. The practitioners who
accomplish the routine tasks may see reality only through the lenses
of their official doctrine; but unless this is completely corrupted it is
likely to be more appropriate to their genuine tasks than the mixture
of misinformation and unconscious prejudice which is provided by
their untutored 'common sense'.18 And the punch-cards and com-
puters, if used rather than abused, can clear away the mountains of
17 A recent e:wnple of this approach, applied in England, was revealed by TIte SUtUlily
Telqrap1l (14 December 1969). For some years the South Eastern Electricity Board
hu secretly operated a 'points' system on CODSUJDel'S, an account losing 'points' every
time a payment is, for whatever reason, delayed. The intended function of the system is
to identify likely defaulters and accelerate the sending of reminden and (fiDally) the
twenty-four-hour notice of cutting off of supplies to than. Naturally, the twenty-four-
hour notice has gone out to people who never consider themselves IS slow payers; and
(because of delays within the postal and accounting systems) some time after the bill
has been paid.
18 I am indebted to Dr. Leslie Walsh for this observation. A confirming instance is
provided by Bob Bailey and Lou Smith, 'Opention Bootstrap', emt" DiMy, 16 (196'7)
(The Fund for the Republic, Santa Barbara, California), 37-44- At their unorthodox
training sch~l in a Los Angeles black ghetto, they had visits from students at the
University of Southern California. 'These are people majoring in race relations who
will probably become the heads of Human Relations Commissions or who will be getting
human relations jobs in industry. They had never, up until the time they came to us,
talked to the very people they're going to be making decisions for. In one ofour sessions
we had a white girl who wu just terrified. She fiDallyadmitted it. She said "I'm scared
to death of Negroes." Somebody asked her why and she said, "Well, they go around
raping people." This is a girl majoring in race relations' (p. 43). One can hope that by
the time she took her degree her formal instruction would have imparted a more soph-
isticated understanding of the problems she would handle in a professional capacity.
354 Science in the Modern World
routine paper so that the genuinely difficult cases can be quickly
recognized and dealt with. But there are real dangers in an extension
of this pseudo-scientific style of handling practical problems. The
artificial worla of the experts and systems analysts, the construction
of whose objects may not have involved much intellect at all, may
be given effective reality in virtue of their institutional and political
power. The solution of a practical problem is, as I pointed out,
determined by the categories in which it is conceived; and if these
determining categories become those which are convenient for a
computer program, we will in that measure live or die by the com-
puters. 19 It is less likely that man "ill be made over into the image
of the punch-card, than that when reality does obtrude itself, it will
do so in an opposition to the imposed systems which is revolutionary
and destructive of them, and of much else besides.

Practical and Technical Problems Compared

The difference between technical and practical problems has been
widely observed in recent years, in the contrast between the great
successes of America and the Soviet Union in the one sort, and their
abysmal failures in the other. The exploration of space is certainly
an astounding technical exploit, which requires inventiveness,
ingenuity, administrative skills, and quality control to an extremely
high degree. Yet these same qualities were insufficient for America
to impose its solution on the conflict in Vietnam, or to resolve the
crisis of its black cities; and the Soviet Union could neither solve
its own problems of management and production, nor permit an
associated state to solve its problems in its own way. The differences
are clear. In the space race, the ultimate purposes were simple, and
established by decree of national leaders: the enhancement of
national prestige. Within this framework, the choice of goals was
not complex, and work could then proceed on purely technical
The artificiality of 'common sense' is well exprased in an aphorism by John Maynard
Keynes in the concluding notes to TIte Gmerfll Tlteory of Employment Inlerest tJntl
Mtmt1 (Macmillan, London, 1936). 'Practical men who believe themselves to be quite
exempt from any intellectual influences are usually the slaves ofsome defunct economist.
Madmen in authority who hear voices in the air are distilling their frenzy from some
academic scribbler of a few years back' (p. 383).
19 The possibility of civilization dying quite suddenly by the computers wu • very
real one in the early 1960s, when American 'nuclear sttategy' relied heavily on computer
simulations of real crises. Sir SoUy Zuckerman was the first eminent military scientist
to expose this lunacy, in 1961; his essay is reprinted IS ch. S, 'Judgment and Control in
Modem Warfare', of SeimtiSlI twl W. (Hamish Hamilton, London, 1966).
 Practical Problems 355
problems; and the harmonizing of their goals with private purposes
of those on the work, was not too difficult. But practical problems
cannot be reduced to a matter of technique, however great the desire
or the financial investment.20
The deepest and most urgent practical problem-situations are
not discovered or invented; they are presented to us, frequently
against our desires, by the processes of human history acting
through time up to the present. The framing and solution ofpractical
problems is at risk of encountering a multiplicity of pitfalls, so that
the purposes served can turn out to be quite different from those
intended. Some at least of these can be avoided by an awareness of
the different phases of the cycle of a practical problem, and its
relation to the associated scientific and technical problems. The
earlier phases involve inquiries of a scientific sort, to determine
what the problem is, and how it might be solved. The first common
pitfall is that the objects in whose terms the inquiry is conducted
are so tightly bound to a particular ideology that the conclusion is
determined before the work begins. But if the inquiry avoids
'theory' and becomes 'empirical', it can encounter the pitfall of
simplifying its objects of inquiry to homogeneous populations de-
fined by classes ofsimple data; then the complexity and contrariness
of the situation, which created the problem situation in the first
place, is lost from view. Thus, to the extent that the conclusions ofthe
inquiry are simple, they are likely to be over-simple; but a conclusion
full of nuance, subtlety, and personal wisdom is inappropriate for
its function as the basis for a decision. Another pitfall of the phase
of 'information and argument' is the imposition of excessively high
standards of adequacy on the work, so that a lengthy and expensive
research programme is undertaken as a preliminary to any decisions.
This may well serve the private purposes of the scientists con-
cerned,21 but the resulting delay may involve costs of its own, not

20 In his article 'Can Technology Replace Social Engineering ?', Bullet;n of tile

Atomit Stimtists, 22 (December 1966), 4-8, Alvin Weinberg argues cogently that
'social problems can be circumvented or at least reduced to less formidable proportions
by the application of the Technological Fix'; but he is fully aware that 'social en-
gineering' cannot be dispensed with.
21 In 'The New Estate', Bullet;n oftile Ato",;t Scimt;sts, 20 (February 1964), 16-19,
Alvin Weinberg gives a good description of this tendency. 'With respect to the style of
intellectuality, I refer to the tendency to deal with difficult teclmo-social problems-
such as the population problem, or aid to the underdeveloped countties, or the lag in
our civilian technology-by deifying research and denying engineering. I have read
 356 Science in the Modem WorltJ
least those of ~ing interpreted by. aggrieved parties as a deliberate
attempt to defer any decision indefinitely.
The function of the conclusions of the phase of enquiry is to
provide a basis for a decision; and this must be capable of execution
as a technical project, by whose means the desired general purposes
may be achieved. The technical project itself must be capable of
being analysed into a multitude of routine tasks, governed by an
hierarchy of decision and control. Further pitfalls are encountered
here; even if the intended functions are appropriate to their ends,
the aggregate of unit tasks, as they are articulated and then con-
trolled by a bureaucracy, may come to be governed by goals which
are contrary to this original function. The operation of overall con-
trol over the phase of execution then constitutes a new practical
problem in itsel~ with its distinct phases of investigation; and the
neglect of this will put the whole practical project in jeopardy. Thus
the inherent complexity of a practical problem, both in its objects
and in the cycle of its solution calls for a diversity of operations,
skills, and approaches even greater than in the case of technical
projects. The application of 'scientific method', not merely as a
crude imitation of the mathematical-experimental sciences but even
in the sense of scholarly inquiry, is insufficient by itself. What is
required is an appreciation of the variety in the nature of the
problems encountered, the difference in their methods, the criteria
of adequacy of their solution, and in their characteristic pitfalls;
and an awareness that there exist some important problems, of
any of these sorts, which are incapable of solution under any
possible circumstances.

'Practical' Aspects o!Technical Problems

We saw earlier how a practical problem can tend, in its phase of
execution, to lose its original defining purposes, and to become just
many reports by presidential panels and panels of our National Academy of Sciences on
a variety of socio-scientific problems. The reports surprisingly often have the same,
basic structure: first, • statement that the problem is difficult and that not enough is
known about it; &ea>nd, that more research and therefore more students and more
fellowships are needed; and finally, that much more government money should be
directed toward research on these matters. All these things are good; but I submit that
in dealing with socio-teclmical problems the university community (that tends to
dominate the panels that prepare such reports) is influenced too much by its ttadition of
scholarship and research and too little by the engineering notion of doing the best one
can with th e available knowledge. It is not that more knowledge is not needed; it is
 Practical Problems 357
another technical project, serving some other external purposes or
perhaps none at all. But there has recendy developed a converse
tendency, in the politically effective awareness of the effects of
technical projects on the welfare of others, independently of their
relation to the organization as customers or users. Thus in America
there have been successful campaigns against the sale of the in-
secticide D.D.T., and in England, an equally successful campaign
against the first proposed site for the third London airport. In each
case, the argument was based on injury to 'third parties', whose
welfare should be taken into account as a final cause in the decision
on the project. Regulation of certain industries on behalf ofa 'public
welfare' of one sort or another has being going on for some time;
what is new in the present situation is the growing strength of the
assumption that every enterprise must be assessed in these more
inclusive terms. This is not due to a sudden increase in the ecological
and social effects of industrial activity, but to the increasing sophis-
tication of politics now that 'welfare' means more than a subsistence
wage for the great majority.
As yet, the tendency to assess technical projects as practical ones
lacks both a coherent ideology and a stable political base. The
traditions on which it draws are scattered and thin, mainly those of
'conservation' of unspoiled areas, and 'preservation' of the countty-
side. In the capitalist societies, the dominant tradition of thought
about the accountability of industrial enterprise has naturally
tended to advocate a minimum of public control; and the opposition
traditions of socialism, based on the sttuggle for the redivision of
wealth, concentrated on social ownership as an automatic guarantee
of democratic control Each of these traditions neglects, in its own
way, the problems ofcontrolling technical projects by considerations
of social welfare. Until very recently, the training of engineers and
scientists, and the rules of their professional institutions, gave no
indication that such practical problems are relevant to their work.
Q!tite suddenly within recent years an awareness has grown, pardy
because the limitations of a technology of affluence are beginning to
be felt, in immediate small problems as well as in large philosophical
ones. 22 The exhaust-fumes of the automobiles of Los Angeles may
simply that in complicated technico-social situations complete knowledge is never at
band, and the engineering approach is as appropriate as is the research approach.'
aa An example of this new awareness is the symposium on 'Technology for Man',
published in Technology tiM Culture, 10 (1969), 1-19.
358 Science in the Modem WorltJ
tum out to be a more genuine harbinger ofthe leading technological
challenges of the future than the magnificent freeways on which
they are generated.
It is important to realize why such problems cannot be either
wished out of existence, or solved either by generalized good will
or general rules. Our analysis of the withering away of ultimate
purposes in a bureaucratic practical project is relevant here, but it
must be supplemented by certain features of technical projects. We
have seen that any successful device must be capable of performing
several functions; and its design must take these into account. In
the case of a mass-produced commodity, these various functions
will be in a rough correspondence to the purposes of different
groups of users; and the device will be commercially successful if
it achieves superiority in a range of functions which appeals to a
large group of prospective purchasers. Thus, the private automobile
can function as commuter transport, suburban transport, long-
range transport, all for different numbers of passengers, as well as
status symbol, fantasy-object, and love-nest; and each particular
model will embody a design compromise, within the constraints of
cost, among such possible functions. However, there are other
devices in which the success of a particular design compromise can-
not be tested retrospectively by the market; as an example, let us
consider a dam.
Among the functions ofa dam, there can be hydro-eleetric power,
water storage, flood control, irrigation, and recreation. Each of these
will serve the purposes of some particular group, through some
benefit conferred by its actions. But the dam will also have 'costs'
or dysfunctions, in the sense that some purposes will be injured by
its effects; thus, it will inevitably be dysfunctional in respect of
habitation and farming of the area to be drowned, as well as of
other sorts of recreation; in its destruction of a natural habitat it
may be dysfunctional for the ecology of the region and for the
sciences dependent on those materials for data; and if material
monuments are also drowned, it will be dysfunctional for the national
culture. Each function or dysfunction is relevant to the interests of
a particular group of individuals; and the groups may have little
or no purposes in common which could serve as the basis for an
acceptable compromise. The dam might even become the focus of
political protest, as when the rural areas to be drowned in the
interests of the city, belong to a different nation; thus the Welsh
 Practical Problems 359
nationalists see the reservoirs on their land as an instrument of
English imperialism.
Special social mechanisms are necessary for the achievement of-a
decision in such cases. Mter public hearings in which injured citizens
vent their aggravation and tame experts display their erudition,
the actual deliberation is done in camera; and the decision, like all
such difficult ones, is probably agreed informally between three
men going down in the lift, or encountering each other in some other
place conducive to relaxed and confidential discussion. In earlier
times, the decision process did not involve so much agony; but this
was because one interest could frequendy ride roughshod over all
others. Things usually got done more quickly, but at the price of
producing ill-designed monstrosities, which ravaged the landscape
and blighted human lives. In our more enlightened age, each special
interest may, if it can organize itself, put its point of view to those
responsible for controlling such projects on behalf of the public.
But the control on these controllers is extremely indirect; although
their decisions are eminendy political, the traditional structures of
political power are not adapted to the exercise of influence on such
decisions. For even if the project is executed by the State, it is the
province ofthe executive rather than the legislature; and the relevant
branch of the executive can always muster enough loyal experts to
keep the odd inquisitive legislator at bay. Hence to the extent that
influence is exercised, it will tend to be done informally, to the benefit
of those who command the channels of informal influence. 23
23 In the United States there is a clear 'life cycle' for regulatory commissions, which
starts with public outcry, and symbolic reassurance, but then is followed by political
quiescence and the conversion of the commission to the furthering of the purposes of
the corporate bodies supposedly being regulated. See Beryl L. Crowe, 'The Tragedy of
the Commons Revisited', Science, 166 (1g6g), 1103-9.
Hints of a deeper understanding of the problem of the conttol of regulatory colDlDis-
sions are provided by Lincoln Steffens, in his chapter 'How Hard it is to Keep Things
Wrong', Autobiogr"plty, pp. 561-«). 'The responsible attorney for a railroad and con-
scientious railroad men have told mc-and convinced me, too-tbat you cannot run a
niIroad without corrupting and ~nttolling government. All discussion of public owner-
ship is foolish; either the State will own and operate the railroads and other public
utilities or these public corporations will "own" and govern the State' (p. 565). But this
form ofunofficial government has its own severe' problems. For as his informant explained
to him, 'The Southern Pacific Railroad and all the companies and interests associated
with us are not rich enough to pay all that politics [i.e. corruption] costs.' Hence» to
keep the machinery of government complaisant, 'We have to let these little sbus
[i.e. the crude, petty, "dirty" grafters] get theirs; we have to sit by and see them run
riot and take risks that risk our interests, too. We can't help it' (p. 567; the order of
quotations is inverted here).
360 Scim&e in tAe Modern World
Political activity on such issues cannot be along traditional lines,
since there is rarely a mass consensus whereby pressure can be
exerted on a national party. It is for this reason, among others, that
Don-violent direct action, conducted as a symbolic gesture bringing
attention to the problem and bringing embarrassment to the bureau-
crats, is becoming popular as a form of practical political activity.
In any debate over a proposed innovation, the most loud and
consistent voice will be that of the group promoting the new device.
At best, they will be honest practitioners of traditional myopic
engineering; their problems are defined in hard, quantitative terms,
and their projected solution may well be the only one possible on the
assumption that the technical and social context of the problem re-
mains unchanged. U But the engineers have recendy been joined by
the aposdes of runaway technology, who when all other arguments
fail have the last refuge of 'progress'. In practice this means that if
an existing device can be 'improved', usually by being made larger,
faster and more expensive, then it is violating a law of nature to
abstain from doing so. The ideology of such 'progress' is given sup-
port by the history of technology, both folk-history and scholarly.
Whereas the history of science has tended to be Whiggish, the
history of technology is quite Hegelian, seeing the efforts of the
past as the unfolding ofthe Idea ofthe Perfect Device of the present
or of the near future. 25 If one restricts 'technology' to the devices
"'In his criticism of a proposal for a 'master drain' to remove brackish ground-water
from the San Joaquin Valley of California, Frank M. Stead, 'Desalting California',
BmJirtnllllnll, 2, NO.5 UUDe 1969), 2-10, says 'But let us not jump to the easy conclusion
that because the drain plan is environmentally UDSOund in the long haul, the engineers
who propose this drain are incompetent or venal I know this is not so. The problem
IteIDI from the shortsighted basic precepts that guide our efforts in the management of
environmental resources in both the public and private sectors. If we want something
better than the San Joaquin Master Dnin, you and I are going to have to change the
environmental management ground rules' (p. :a).
as The strength of this conception of 'technology' is well illustrated by a story of
Lewis Mumford: 'Unfortunately, so firmly were the nineteenth-century conceptions
committed to the notion of man u primarily IumuJ f.m, the tool-maker, rather than
__ ",pints, the miDd-maker, that, u you know, the first discovery of the art of the
AItamira caves was dismissed as a hoa because the leading pa1eo-ethnologists would
DOt admittbat the Ice Age hunters whose weapons and tools they had recendy discovered
could have had either the leisure or the mental inclination to produce art-not aude
forms but images that showed powers of observation and abstraction of a high order.'
See 'Technics and the Nature of Man', TttimolDO "lUI Cult"t, 7 (1g66), p. 309.
For an example of I reoognition of the limits of this historiographical tradition, we may
cite T. Ie. Derry and T. I. Williams, .If Sluwt History ofTtt"noloO (Clarendon Press,
ODonI, 1960), Epilogue, p. 710: 'From this standpoint (the happiness ofthe individual)
the greatest indisputable benefits of modern technology are perhaps those conferred by
 Practical Problems 361
created by the mechanical and civil engineering of the last few
centuries, and concentrates on a simple sequence of advances, one
can render a plausible story along these lines. The time-lag
between advances can be explained by unfavourable technical or
commercial circumstances, and the delays in diffusion to other
places can be similarly explained in terms of their 'backwardness'
in some respect. The moral of 'inevitable progress' always
emerges from such an history, partly because of its inherent
plausibility and partly because it was built into the study in the
first place.
We have already seen that a new device is successful only if it
can perform some functions whereby certain purposes are served.
In the case of a device put on the market, both consumers and pro-
ducers must derive benefit from its production and sale; and to-
gether they must be able to override the objections of those who are
injured by the innovation. These latter can include those whose
interests are threatened by the displacement of an existing system of
devices (either capitalists or workmen), and those who will suffer
incidental harm from the adoption of the new device. Over the last
couple of centuries, the capitalist system has been weH suited to the
fostering of innovations and the overcoming of resistance, in manu-
facturing and agriculture. With the passing ofthe guilds of Medieval
times, and of the Royal patents for monopolies of a later period, the
protection of capital against innovation was lost. The artisans and
workmen whose work would be degraded or displaced by innova-
tion could be suppressed by economic and political techniques.
And those who were injured by an innovation could either move away
if they had the money, or remain and suffer if they didn't. Outside
the manufacturing industry and agriculture of the nineteenth and
early twentieth centuries, the constraints on innovation were not
so easily swept away. House-construction, for example, is only now
moving forward from the materials and organization developed
millennia ago; and the 'technology of the home' can be seen as far
more complex cultural phenomenon than might appear from a
casual glance at the latest refrigerators and televisions. 26 Also, as it
becomes increasingly difficult to escape from the degradation and
bnnches which the present Short HistMy has Jacked space to emphasise u they deserve,
namely the revolutionary changes in medicine and surgery.'
26 A case study in the complexities of the diffusion of innovations is given by A.
Ravetz, 'The Victorian Coal Kitchen and its Reformers', Vittoritm Stllllies, II (1g68),
435-60·
 362 Science in the Modern World
squalor of mindless industrialism, sections of the articulate and
educated classes have joined the protests.
The inadequacies of the simplistic 'progressive' View of techno-
logical development can be seen most clearly in the case of the auto-
mobile. On the one hand, it is possible to trace a development from
the first primitive machines, through a sequence of models of in-
creased speed, comfort, reliability and safety. Yet even this internal
history is not one of steady, autonomous improvement. The safety
of automobiles has received serious attention only recendy, in re-
sponse to campaigns waged by independent critics. And the speed
of automobiles has stopped increasing; although they can be made
to ride comfortably at some ninety miles an hour, in all the countries
where the automobile is a standard piece of domestic equipment it is
recognized that at more than seventy the weapon becomes too lethal.
Also, the external constraints on the diffusion of the automobile are
now becoming recognized. For years, highway engineers attacked
the problem of traffic congestion by building bigger roads; but
recendy they have become aware ofthe; principle that any convenient
automobile channel soon becomes clogged beyond capacity. Finally,
the damage done to the atmosphere, and to the social environment
by the present stage of progress in private transport, is recognized
as serious and urgent. Thus, in this most important sector of the
modern economy, the belief in the necessity of simple 'progress',
and the duty to participate in it, has collapsed in the face of the
realities of technology in its social and ecological context. 27
The problem of achieving effective democratic control on the
decisions oftechnological innovation is beset by the many difficulties
we have seen: the combination of contrary purposes, with bureau-
cratic operation, and the different ideologies of the various groups
in~olved. Yet if this problem is not solved, our social lives will
inevitably come to be ruled to an increasing extent, by blundering
bureaucrats and experts. Worse yet, this will be recognized as a new
form of oppression, which can secure acquiescence neither by the
possession of direct police powers by the ·decision-makers, nor by
a7 An illuminating survey of the intricacy ofthe 'welfare' aspects of technical projects
is given by T''''"ltWtl#itm "lUI CtmlllUlllity VilifieS (Special Report lOS, Highways
Research Board; National Academy of Sciences, Washington, 1969). The projects
disaJaed there are the urban motorways, designed for the convenience of the suburban
whites, and causing distress and disruption to the urban poor, mostly black. Only after
the pat urban riots did the planning engineers discover that there is a genuine conftiet
of values, not easily reducible to quantitative terms for a systems analysis.
 Prll&tica/ Problems 363
the legitimacy derived from a semblance of answerability to the
population. Such problems have already appeared explicidy in the
Socialist countries, where the bureaucracy is quite obttusive; and
they can be interpreted as the cause of much ofthe current 'student'
unrest in the capitalist societies as well. 28 The first ideas towards a
solution, which involve drastic changes in our inherited conceptions
of democratic politics, have already been put forward. To his great
credit, Mr. A. Wedgwood Benn made a positive contribution to
this long-term problem even when he was, as Minister of Technol-
ogy, enmeshed in the destructive practicalities of the present. 29 But
the full solution of this very deep political problem is a task for the
decades to come; and on it may well depend the survival of our
civilization.
21 Sec Hannah Arendt, 0" ViDlmee (Harcourt, Brace & World, New York, 1970).
The dominion over man is today exercised by I 'bureaucracy or the rule of an intricate
system of bureaus in which no man, neither one nor the best, neither the few nor the
many, can be held responsible, and which properly could be called rule by Nobody. If
we identify tyranny u government that is not held to give an account of itsel( rule by
Nobody is clearly the most tyrannical of all, since there is DO one left who could even
be wed to answer for what is being done. It is this state of af£ain, making it impossible
to localise responsibility and to identify the enemy, that is among the most potent causes
of the current world-wide rebellious unrest, its chaotic nature, and its dangerous ten-
dency to get out of control and to run amuck.' (Q!loted from • review by Fred J. Cook,
Tile Nlllitm (6 April 1970), 606.)
29 Nigel Calder, TetAnolDlis (MacGibbon & Kee, London, 19(9) discusses the Czech
investigations of this problem before 1968, along with the proposals of Mr. Wedgwood
Benn and various Americans, in ch. 17, 'Democracy oCthe Second Kind'.
 14
IMMATURE AND INEFFECTIVE
FIELDS OF INQUIRY

OvER the centuries, philosophical inquiry into the nature of

scientific knowledge has concentrated its attention on fields which
were already matured and effective. These provided an example of
what ~uld be achieved through the study of"the natural world; and
it was plausible to assume that they could also serve as a model for
the methods of other, less fortunate fields. But such fields are a
minority, even within the study of the natural world. To ignore the
others is to present a distorted view of scientific inquiry, and to
mislead scientists who are trying to develop their fields to the same
state ofeffectiveness as those whose success is universally recognized.
For when the fully matured fields are taken as the normal, the others
are implicidy classed as abnormal, and their condition is a matter for
concern and even shame. Attempts to improve them by a mechanical
imitation of the methods of the paradigm successful disciplines, in
our times the matured mathematical-experimental fields of natural
science, can do little good and much harm.
In order to analyse this particular pathology of scientific inquiry,
we will need to use a more sophisticated battery ofconcepts than has
been developed hitherto. The dominant traditions in the philosophy
of science have assumed a simple dichotomy between genuine
science, either already matured or capable of maturation by the
application ofsome straightforward techniques, and pseudo-science,
doomed to ineffectiveness by a misconception of, its objects of
inquiry and its methods. The list of modern pseudo-sciences was
fixed during the seventeenth century; it included astrology, alchemy,
geomancy (divination by local features of the earth's surface), and
phytognomy (ascription of medicinal powers to plants by their
suggestive shapes, as developed in folk traditions). Its characteristic
was the assumption of sense and meaning in the ordinary objects of
 Immature and Ineffective Fields ofI"'luiry 365
the external world, and its methods involved a measure of dialogue
with them. By conttast, genuine science assumed no human proper-
ties in its objects, and proceeded by some mixture ofobservation and
reason.
This simple distinction had a very important ideological function
during several centuries, while 'science' was still involved in a
struggle against 'dogma and superstition' for authority in pronounc-
ing on the natural world. The problems of the philosophy' of science
were organized around the specification ofthose methods that would
yield genuine scientific knowledge, and then the character of that
knowledge as a system yielding truth or its acceptable substitute.
The appellation 'pseudo-science' conveyed the deliberate connota-
tion that such fields were not merely ineffective, but also deceitful,
and pernicious, standing for darkness against enlightenment.
Because the historical associations of the term 'pseudo-science'
are still so strong, its mention inevitably suggests a deep and blanket
condemnation of a field so described. I shall therefore avoid it and
instead employ terms focussing on one or another aspect ofthe state
of ineffectiveness. We recall that we could define this state, in terms
ofthe absence of 'facts', a condition caused by the absence of criteria
ofadequacy appropriate for the detection and avoidance of pitfalls in
research. A more commonly used term for this condition is 'imma-
turity', and I shall occasionally use it as well. But this has the
connotation of denying the complete ineffectiveness of the field, and
also promising a development towards matured and effective state;
and as such may well be misleading. Indeed, one of the most
delicate problems of the history of science is the assessment of a
field now successful when it was in an earlier, ineffective state. The
historian must decide whether it was then 'immature', containing
within itself the needs of later successes; or whether the subsequent
innovations were so deep as to involve the transformation and
rejection ofthe essential features ofthe inquiry as then practised. On
this assessment will depend the evaluation of the field as it was then,
and the consequent practical decision on its inclusion in programmes
ofteaching or research. The most famous instance ofsuch a problem
is the field of mechanics in the later Middle Ages, where some
formal 'anticipations' of Galileo's results are to be found; and the
significance ofthese for the development ofclassical mechanics from
Galileo onwards is still under debate.
The problem of assessment is even more acute in the case of a
 366 Scim&e i" the Moder" World
field which is currendy ineffective, but whose leaders claim it to be
on the point of emerging from immaturity. This amounts to a claim
that the objects of inquiry and their associated methods are approp-
riate for the development of the field; while those who disagree
would consider further investment in the field as it stands to be futile.
The correct assessment can only be made retrospectively, and even
then mvolves deep difficulties. Because 'immaturity' and 'ineffec-
tiveness', although quite distinct in their implications, are connected
by so many nuances as well as being difficult to distinguish in
practice, I shall use both terms in any discussion nearly inter-
changeably.
At the present time, the disciplines that present the most obvious
evidence of ineffectiveness or at least immaturity, are those which
attempt to study human behaviour in the style of the mathematical-
experimental natural sciences. But it is important to keep in mind
that each of the natural sciences in the past, and special fields within
them at the present, have had and do have this character. Earlier
generations of historians of science tended to make a sharp distinc-
tion between the darkness of the 'prehistory' ofa subject (perhaps as
lightened by one or a few great men speaking truth before their time),
suddenly transformed into maturity; hence synthetic accounts of the
evolution of scientific disciplines in these terms are rare. For
medicine, it is possible to interpret its history from its beginnings to
the end of the nineteenth century, as reactions to the insoluble and
yet pressing problems posed by its being in a state ofineffectiveness. I
As we shall see, the difficulties of working in an immature or
ineffective field are serious and manifold. Added to the basic diffi-
culties of hying to do research in a field where the pitfalls are still
unidentified, there are the social constraints forced by the pretence
of maturity. The situation becomes worse when an immature or
ineffective field is enlisted in the work of resolution ofsome practical
problem. In such an uncontrolled and perhaps uncontrollable con-
text, where facts are few and political passions many, the relevant
immature field functions to a great extent as a 'folk-sciehce'. This is
a body of accepted knowledge whose function is not to provide the
basis for further advance, but to offer comfort and reassurance to
some body of believers. Our discussion of the applications of
immature or ineffective fields depends heavily on this idea of 'folk-
science', and so two sections on it are inserted in the chapter. We
I For such an interpretation, see R. H. Shryock, Tile DewID,.", 0/MoilmJ M,tlidne.
 Immature and Ineffective Fieltls ofInlJuiry 367
should observe that not every folk-science is associated with an
immature academic discipline; some have none, while others are
associated with fully matured disciplines. Conversely, not every
immature or ineffective field is associated with a popular folk-
science; but the presence of such an association can explain an
otherwise incomprehensible popular trust in a discipline of this sort
when it occurs.
Using these materials, we can study the state and causes of the
condition of ineffectiveness, and then see in what ways knowledge
can come to be achieved in spite of it. We then move on to the social
aspects of inquiry in such fields, discussing the reasons for the
pretence of maturity and its effects on research and teaching.
Passing to the relations of such disciplines with the outer world, we
consider the really complicated situations that arise when they are
developed in connection with technical and practical problems, and
also entangled with ideologically-sensitive folk-sciences. Since the
most important applications of 'science' to the problems of society
have just these features, an appreciation of these difficulties and
pitfalls is important for the proper use of science in the modern
world.
Ineffectiveness-the Absence ofFacts
In a matured field, the processes ofselecting and transforming the
results of research for the achievement of facts can proceed in such
an orderly manner that the nature ofthe 'facts', as stable products of
these operations, can pass unnoticed. But here as elsewhere, the
pathological state an provide clues for understanding the normal.
The indubitable and public symptom of ineffectiveness of a field is
the absence of facts, in the sense in which I have defined them.
Moreover, in such fields the deficiency is most obvious at the level of
elementary teaching. There, unlike in a matured field, the students
do not encounter a collection ofstandardized materials, presented in
a digestible form, and utterly reliable and incontrovertible in them-
selves in spite of the slighdy different guises in which they are
presented in different books or lecture-courses. By contrast, in the
ineffective or immature field, the student is presented with one out
of several sets of supposedly basic materials, and can discover other
sets by reading textbooks not on the recommended list. These
materials themselves consist of intuitive generalities dressed up as
empirical laws, and insecure theoretical speculations masquerading
368 Science in the Modern World
as fundamental explanations.:& The lay public sees the same phenome-
non from a slightly different aspect. Watching the activity of a field
over a period ofyears, one does not witness the steady cumulation of
new facts, ~rhaps superseding but never completely destroying the
old. Instead, there is a succession of leading schools, each with a
manifesto which is more impressive than its achievements, and each
passing into obscurity when its tum on the stage is over. 3
The ineffectiveness of a field, as revealed by the absence of facts,
is a trying state of affairs in the best of circumstances. There are
many natural reasons why the leaders of such a field should try to
pretend that it is otherwise, or to claim and hope that the condition
is easily remedied. Most such claims, and the research strategies
organized around them, involve a concenttation on one aspect of the
work which is seen to be deficient, and whose improvement should
produce full maturity. Heroic attempts have been made to amass
empirical data; to apply mathematical and computational tools for
the production of information; to construct elaborate systems
representing objects by symbols and manipulating them in a formal
argument; and finally to develop methods and a methodology appro-
priate for the discipline. In each case, the attempt is to reproduce
what is believed to be the crucial feature of an established science,
where this is learned more from philosophers ofscience than from the
successful practitioners themselves. But almost all these one-sided
efforts fail utterly. Those which survive over time, through the
sttength of the founder of a school, and institutional stability, are
likely to produce results which become elaborate and refined to the
point of being grotesque. For the condition of ineffectiveness is not
2 This instabililty in the ineffective sciences was of crucial importance in the develop-
ment of the ideas of Thomas Kuhn. In the Strw"", ofStimtiji& Rt'f)(JlulitmS, he reports
on his experience at the Center for Advanced Studies in the Behaviounl Sciences,
' ... I was struck by the number and extent of the overt disagreements between social
scientists about the nature oflegitimate scientific problems and methods. Both history and
acquaintance made me doubt that pnctitionen of the Datura1 sciences possess firmer or
more permanent answers to such questions than their colleagues in social science. Yet,
somehow, the pnctice of astronomy, physics, chemistry or biology, normally fails to
evoke the controversies over fundamentals that often seem endemic among, say, psy-
chologists or sociologists' (p. x).
3 In his Editor's Introduction to HfIIUlIHJoi of MtHltm SotioloD (Rand McNally,
Chicago, 1966), R. E. L. Farris discusses the early history of American sociology which
proceeded from manifestoes, through the establishment of formal institutioDS, to
methodological debates. The process started in 1893; he says 'In a sense, sociology
became organized before it began to exist.' Only some thirty years later did viable
results begin to appear, in the work ofThomas and Znaniecki, and the latter could hardly
be counted u 'American' at that time at least (pp. 26-7).
 Immature and Ineffective Fields ofInlJu;,y 369
an accidental deficiency in some component ofthe materials ofa field,
but is a systematic weakness in those materials and in the social
activity whereby they are produced. Projects for immediate and
radical improvement in a field must necessarily assume a simplicity
of the means to their end; but they then become similar to those for
radical reform or revolution in society, requiring the same qualities
of dedication or fanaticism in their leaders, and incurring the same
risks of patent failure and illusory success. 4
In analysing the absence of facts, we naturally consider first those
materials which should, but do not, achieve this status: the con-
clusions of arguments, and the evidence, information, and data on
which they are based. The failure of nearly all of these to survive
even in the short run, is an indication that most of the work of
investigating problems is vitiated by pitfalls, encountered sooner or
later in the work. The results ofresearch are generally weak, or even
vacuous. This condition prevails even in fields where the leaders and
their associates spare nothing in their endeavours; but the absence of
a body of appropriate methods of inquiry nullifies their efforts. For
it is through such methods, ranging from the techniques of produc-
tion of data, to the judgements of adequacy on an argument, that
pitfalls are identified, and ways around them are charted. Because of
the subdety and sophistication of scientific inquiry, these methods
are a craft knowledge, built up by successful experience. But an
ineffective or immature field has no such experience; and so the
improvement of its methods is not a straightforward operation. 5
4 As an influential English example of the many attempts in this direction, we may cite
Baroness Barbara Wootton, TtSI.mmt for lilt Soti.l Stittl&ts (Allen It Unwin, London,
1950). Learning from her sources in the philosophy of science that science consists of
observation, hypothesis-testing, and measurement, and thus being spared any intima-
tion of the artificiality of the objects of scientific inquiry, she could conclude that simple
diligence and care could bring the social sciences to the same state as physics. She wu
troubled by one sign of immaturity: that the classics of political and social theory do
not become obsolete in teaching. Her exp1aDations are not entirely CODSistent: that
these practices are retained through a prejudice against the scientific status ofthese discip-
lines; and that 'perhaps the need adequately to fill a curriculum has something to do with
this' (pp. 29-30). In fairness to her, we should record that in her soeWsnme, tMUl Sotitll
P.,MWD (Allen It Unwin, London, 1959), derived from her pnctical experience, she
recognizes the pitfalls to be encountered in the simplistic application of a science (in
this case psychiatry) to social problems.
5 The absence of appropriate criteria of adequacy in the rapidly-growing field of
'technological forecasting' is shown clearly in Martin Shubik, 'Processing the Future',
Sn"",, 166(1969), 1257-8,areviewofR. U. Ayres, Tttlmowfi&1Il PortttUti"g.1UI Ltmg-
R.",t PltI"";,,, (McGraw-Hill, 19(9). Commenting on the author's enthusiasm for
computer analysis of 'hundreds of thousands' of sceoarios, he says, 'Tons of pages of
370 Scim&e in the Modern World
The weaknesses in the social aspects of inquiry also conttibute to
the self-perpetuating condition of ineffectiveness. The mechanisms
for the processing of results, and for the exercise of quality control,
cannot be sttonger than the materials on which they operate. For
social reasons it is necessary to give the formal authenticity of
publication to masses of results which are very weak; and so the
effective standards of quality cannot meet those of a matured field.
Because ofthe rapid succession ofseparate schools, each with its own
objects ofinquiry and principles ofmethod, there is little opportunity
for results of potentially high quality to survive and become estab-
lished as facts. And in this unstable and frequently false social
situation, the mechanisms for the control of quality and the
maintenance of scientific integrity at the highest levels do not
exist. 6
To bring a field out of an immature state requires first of all a
sttengthening of its objects and methods; and this requires the
endeavours of a man of exceptional talent and dedication. Such a
man will have a strongly marked personal style, both in the scientific
and social aspects of his work, and probably a personal commitment
to a goal which is deeper than the mere establishment of positive
knowledge in the field. Hence, as we have discussed in connection
with style, a field in a phase of maturation will still bear certain
marks of its earlier history, both in its social organization and
doctrines. Also, the existence of a body of established facts is not in
itself sufficient to ensure the proper operation of the social mechan-
isms of direction and quality control. A field can enter a degenerate
and stagnant state, where its members and leaders are incapable or
unwilling to devote their efforts to its further advancement, but are
content to produce mediocre or shoddy work while basking in the
reflected glory of their predecessors. But this problem has already
computer output are in general • sign of • badly understood job. Most simulations have
• mue inversely related to the fourth power of the quantity of computer output.' For
corrective reading, he prescribes G. A. Miller, 'The Magical Number Seven, Plus or
Minus Two-Some Limits on our Capacity for Processing Infonnation',:/ounull ofllle
A,tnlStie,,1 Soany ofAmen,,,, 22 (1950), 725-3°. Of course, in this case as in others, the
Americans compensate for their invention ofteehnological monsttosities by their ready
wit in describing them; thus Shubik's point is expressed by the acronym GIGO-
Garbage In, Garbage Out-for vacuous computing.
6 Paul E. Meehl, 'Theory Testing in Psychology and Physics: I Methodological
Pan dos', PlalD.,'y'fSlime" 34 (196'7), 103-15, shows how sophisticated significance
tests give. spurious 'confirmation' of hypotheses; even random data would pass the
tests h .If the time. In spite ofbis distinpUshed position in American psychology, he has
been UDable to prevent the widespread adoption of such lax standards of adequacy.
 Immature and Ineffective Fields ofI"'luiry 37 1
been discussed in connection with ethics and quality control, and we
can concenttate here on genuinely immature fields.
Nor should it be imagined that the problems discussed here are
restricted in their relevance to entire disciplines which are recog-
nizably ineffective or immature. We have already seen in connection
with methods ofresearch, that any major advance in science involves
an innovation, which temporarily at least is not capable of assess-
ment by t.'te existing criteria for work in its field. Hence whenever a
discipline extends a deep salient into the unknown, that point will
have the characteristic features of immaturity to a greater or less
extent. 7 In a mature and healthy field, the social processes of
selection and transformation of results will bring back from the
research front, for more general use, materials that are 'matured' in
the passage, and add them to its stock of facts. By conttast, in an
ineffective or immature field, there is nothing there behind the front
lines; and finally, a sign of stagnation and senescence in a field is
when it is all completely matured and safe, hence completely safe
and boring.
Knowledge outside the Matured Mathematical-Experimental Sciences
One reason for the difficulties ofimmature and ineffective fields is
that their model of genuine science is a very specialized one, which
7 A most illuminating description of the immature state of • new field in physics is
given by D. C. Montgomery and D. A. Tidman, Plimu IGnetie Tile.,:! (McGnw-Hill,
1964), preface, p. vi. On theory and experiment: 'there is so far very little overlap in
plasma physics between accurately measurable and accurately calculable quantities. The
experimentalist becomes entangled with situation-dependent panmeters-percentage of
impurities, details of vacuum systems-while the theoretician concentrates on one-
dimensional, unbounded, and otherwise highly idealized situations, so that opportunities
for comparing results have been regrettably few. The indications of the last two years
have been hopeful, however, and this disjointedness may well be on its way out.' On the
theory itself, we read: 'The tendency in kinetic theory, since long before the days of
plasma physics, has been to confuse what is assumed with what is proved. This has not
been disastrous in the area of classical molecular gas dyuamics, since there wu always I
steadily accumulating wealth oCexperimental data to provide ballast Cor the theoretician.
In plasma physics, however, most of even the simplest theoretical predictions go quan-
titatively unverified experimentally, and there has been I profound and unfortuDate
separation of what has been calculable and what has been measurable. It is • truism of
modem science that the rewards (or writing papers are greater than those (or reading
them. Plasma physics, perhaps more than any other branch, has sufFered from undue
attention to this dictum and has generated, in so doing, • literature so large, unwieldy,
and hastily put together that the newcomer to the field finds that it is nearly impossible
to verify the truth or falsity ofany given theory to his own satisfaction.t I am indebted to
on~ of the authors, D. C. Montgomery, (or telling me o( this source, and (or fruitful
discussions of these problems, when we were both in Utrecht in 1964-5.
 372 Scim&e in IIJe Modern World
may be quite inappropriate to their own tasks. To understand the
special problems of immature fields, and the possibilities for their
development, we can derive guidance from the history of scientific
inquiry in the period before the rise to dominance of 'positive
science'. Then, there was a clear distinction between two sorts of
inquiry, 'history' and 'philosophy'; and they were in turn distin-
guished from 'art'. The former meant description, not merely of past
times, but of any class of objects; and 'natural history' still survives
as the tide of some long-established museums. 'Philosophy', on the
other hand, meant reflection and explanation, as applied to any
problem. Thus Dalton's atomic theory was announced in his 'New
System of Chemical Philosophy', and what we now call 'physics'
was known as 'natural philosophy' in England through the nine-
teenth century. Each type ofinquiry had its own criteria ofadequacy,
the one emphasizing faithful and comprehensive accounts, and the
other coherence ofargument. Ofcourse, there was plenty ofvariation
in these, over time and place, and many were the debates in which
each side accused the other of being 'unphilosophical'.
Between the two sorts of inquiry, 'history' and 'philosophy', they
omitted something distinctive of positive science, and also main-
tained something that has since been lost. For the positivist ideal ofa
system of advancing knowledge, in which bold speculation proceeds
in a tight connection with particular facts from experience, was not
the ruling one. Newton's work is a partial exception to this generali-
zation, but even he had recourse to 'Q!1eries' when he published his
views on the deeper problems of what we now call physics. On the
other hand, the philosophy of nature was, in subject matter and in
spirit, continuous with philosophy in general; the difference between
them in the nature of the evidence and the conclusiveness of the
arguments was one of degree only. Although it would not be true to
say that all the productions in natural history and natural philosophy
were good of their kind, these conceptions of the nature of the
inquiry were appropriate both to the state of knowledge, and to the
conditions of social organization of the fields. The activity of
'research' requires a certain degree of maturity of the field, and also
elaborated social institutions, if it is to be effective. 8 When the study
• It is only very recendy that the term 'natunl philosopher' was driven out by
'scientist'; in England the conquest was not complete until the present century. See
S. Ross, 'Scientist, the story of. Word', .A_Is 0/ Seitrlte, 18 (1g62; published 1964),
65-86.
 Immature and Ineffective Fields ofInlJu;ry 373
of nature was pursued largely by isolated individuals, in fields far
from maturity, it was as well that each one could pursue the sort of
problem that has a lengthy cycle of development and can be worked
in an individual style.
The recognition of 'art' as a category distinct from inquiry could
also contribute to a better self-understanding of immature fields.
The ttaditional term was not restricted to subliterate handicrafts;
for Aristode, it extended to the set ofprinciples defining the methods
of any class of tasks. Arts could be 'liberal' as well as 'mechanical';
and could involve a sophisticated (and genuine) scientific component,
as the Renaissance 'art of navigation'. Now, a literate art would
naturally be based partly on the 'history' of its objects, and be
informed to some extent by a 'philosophy' of its principles. On
occasion, a rising art would develop a 'philosophy', with the function
ofenhancing its prestige; such was the case with architecture during
the Italian Renaissance. 9 But these related inquiries would be
ancillary and incidental to its real work, whose strength and success
was independent of theirs. 10
In the modern situation, we have many examples of successful
special 'arts' arising out of inquiries in immature or even nascent
fields. These will usually be methods for .solving technical problems
which, although very restricted in comparison with the goals of the
field as a whole, are quite useful. It is known from experience that
the methods work, and a body ofgenuine craft skills can be developed
for the tasks; even though the theory on which the method was first
based may have been discredited, and no scientifically satisfactory
explanation has been produced. It is a natural error to cite such arts
as evidence for the maturity of the field, as if they arose as applica-
tions of a solidly established body of fact. I I But it would be more
honest and more fruitful to accept such arts as those points where the
9 This is not to say that those who proposed the 'philosophy' were necessarily ccmscious
of this social function. In the case of the Renaiaance theory of harmonious proportion,
the motive could well have been the exteDsion of the Platonic idel1 of knowledge to aD
areas of experience. See R. Wittkower, Al-e.teaw"l Prinei,les in tile Ate ofH"""";m,
2nd ed. (Alec Tiranti, London, 1967).
10 Even in the case of Renaiaance an:hitecture, the greatest architects were not the
greatest theoreticians; and greatness was achieved by aeation within I style ntber than
by working out exercises in the application of its theory.
II Thus, E. G. Boring, .A History of ufJlrifllllltlll Psye1uJwD, md ed., reviews his
usessment of 1929, that American psychology wultill immature; and cites its quanti-
tative growth and its applications as evidence for I satisfactory state of maturity (pp. 743-
743).
13-S.K.s.P•
374 Science in the Modem World
discipline has somehow made an effective contact with the external
world, and with some humility to consider how this might have
occurred.
Ifthe leaders ofineffective and immature fields at the present time
were able to recognize their condition publicly, and to institute
appropriate methods, then this distinction between 'history',
'philosophy', and 'art' might be conducive to a more healthy atmo-
sphere. Where the objects of inquiry have but a tenuous relation to
the real things and events they purport to describe, and are them-
selves ill-formed and unstable, an isolated investigation devoted to a
supposedly 'empirical' test of some hypothesis about their relations,
is highly unlikely to yield worthwhile results. As an alternative to
such 'research', we can imagine a sort of 'history', conducted in a
disciplined fashion and using all appropriate tools, whose objects of
inquiry are those of a trained common sense, and which has a less
formalized, and correspondingly more extensive and perhaps deeper,
contact with its sources. The discipline which we now call history,
works in this fashion on the traces of human activity in the past; and
it can offer tested methods, including criteria of adequacy, to this
sort of work. At the other extreme, 'theory' could be recognized as a
sort of 'philosophy', and could be governed to some extent by the
methods of that discipline. It would then not be a cause of surprise
or shame that effective new insights come only very rarely, and that
they are only slighdy developed by the subsequent efforts of lesser
men; and it could be recognized that an esssential part of a genuine
education in the discipline is a dialogue with its great masters.
Finally, the successful 'arts' could be recognized as the most
genuine tested experience of the field, and studied and developed as
such. Each of these three sorts of inquiry could eventually yield
knowledge of its characteristic sort and, in their interaction, offer
mutual criticism and support.
It might be asked, what would then happen to the middle ground
between 'history' and 'philosophy', which constitutes the distinctive
achievement of positive science: the framing of general laws,
universal in application, which are tested and confirmed by con-
trolled experience? The question is not relevant to this discussion,
for here we are considering just those fields where this does not
exist; such laws are one variety of facts. It is even doubtful that this
should be the goal of the development of every field of disciplined
inquiry; but in any event the simplistic pursuit of that goal has
 Immature and Ineffective Fields ofInlJuiry 375
clearly failed to bring results in many cases. Even with the enriched
conception of the work in an immature field suggested here, the
'quality cut-off point' would still be higher than in those where
routine technician's work is sufficient for the production of com-
petent results. For a good history requires personal insight into the
phenomena, and philosophy is a very demanding inquiry. But the
resulting reorganization of the materials of the field and of the con-
ception of the work could save many a talented recruit from years of
futility.
Moreover, it is not necessary for a discipline to be fully 'positive'
in the sense of imitating physics for it to make a contribution to the
advancement of human knowledge. Facts of the ordinary sort can be
achieved in history by processes which are the same in principle as
those we have described for experimental natural science, although'
not so closely controlled in all their operations. In philosophy too,
there are certain problems and solutions which are widely accepted
as fundamental, and this 'classic' doctrine can be and is successfully
taught to students. However, the eventual products of the processes
of selecting and transforming 'facts' of these sorts have a different
form from that of knowledge in the matured mathematical and
experimental sciences. A very deep result in these other disciplines
will generally not survive as a simple, impersonal, and apparendy
elementary and indubitable assertion. It is more likely to appear as
an aphorism: an expression of a deep personal understanding of its
objects, in a condensed and communicable form. When the ideal of
knowledge is that of an impersonal and rigid system, aphorisms are
scarcely recognized as bearers of knowledge; and this has been the
situation in the philosophy of science, and in the dominant currents
of epistemology, since the seventeenth century. But for Francis
Bacon, one of the pioneers of the 'new philosophy', who had deep
roots in an older, humanistic tradition, knowledge was distilled
rather than deduced, and aphorisms were correspondingly impor-
tant. He believed that the deepest and most general principles of
nature would be expressed as aphorisms; and his Novum Organum
was itself organized as a series of aphorisms. He contrasted the
fruitful and living character of aphoristic knowledge, to the rigid
and frequendy sterile character of deductive systems. In this, he
doubdess had in mind the standardised versions of Aristotelian
doctrine that were taught in the universities. But ofequal importance
to him was theology, of which he said that a collection of aphorisms
376 Science in the Modem World
from English sermons of the preceding half-century would be 'the
best work in divinity which had been written since the aposdes'
times'. 12
Thus, to the extent that a discipline differs from the state of being
matured mathematical or experimental science, its permanent know-
ledge, in whose terms its facts are organized, will be embodied in
aphorisms rather than in impersonal universal laws. This applies both
to immature fields and to those whose materials do not enable the
establishment of evidence from rigorously controlled experience, or
of arguments with a complex and rigid structure. And at the far
extreme, where a craft knowledge finds verbal expression, all its
principles will be expressed as aphorisms. When compared with
'scientific' knowledge, aphorisms suffer from several defects. Since
they are not the conclusions of a tighdy-sttuctured argument, they
are incapable of modification through criticisms of particular
IS It seems likely that in his earliest years, Bacon hoped that personal wisdom could
be obtained directly from a collection of aphorisms; a record of this endeavour is pre-
served in his Colows ofGootl .fIIl Evil (Woris, vii, 75-92), and in the De Augment;s
StitllliMuM, Book VI, ch. 3 (translation in Woris, iv, 473-91). Although he apologized
(or these in the earlier piece, and relcpted them to the section on transmission of know-
ledge in the later book, there is evidence that he considered aphorisms to carry the
deepest hUths about God's creation. For, in ch. 1 o( Book III of the De Augment's, he
pvc a tentative list of 'uioms' which would CODStitute Pltilosoplti. Pri"", or S.pimt;.;
and showed that they hold hUe in all disciplines of nature and man. Indeed, he con-
sidered them u 'plainly the same footsteps of nature treading or printing upon different
subjects and matters'. An example of this class of axioms is : 'Putrefaction is more con-
tagious before than after Maturity', hUe for 'physics' (1 medicine) and for morality
(translation pp. 337-9).
Bacon's remarks on aphorisms as a source of knowledge are reiterated through his
writinp, starting with the AtlWW:emnJt ofu.";,,,( Woris, iii, 292). There he condemned
the 'over-early and peremptory reduction of knowledge into arts and methods', using
the argument that things which come to a perfect shape too soon seldom grow further.
In the COIiI.'. el Vi. of a few years later (Woris, iii, 593-4; translated in Farrington,
Tile Pltilosoplty ofF,."as BM"", Liverpool, 1964, p. 75), he used the precedent of the
IDCicnt Greek philosophers (now ca1Ied the 'presocratics'). He praised the use of
aphorisms (or giving a plain account of things discovered, and being silent where there
wu nothing discovered; and also for stimulating the mind to judgement and discovery.
Part of this passage was taken over into Aphorism 86 of Book 1 of the Novum Org."um
(translated in Woris, iv, 85); there he added the comment that the pretence of sys-
tematic knowledge in incomplete sciences, results from the 'craft and artifices' of those
who handle and transmit them; with a resulting sterility in those sciences.
The remarks on divinity are near the end of the AtlN1I&emnJt ofU"';"g (Woris, iii,
487-8). It must not be thought that Bacon was opposed to all systematizations. The con-
solidation and reform of all the branches of English law was of great concern to him
through his entire public career. His shon paper of 1616, defining a project to that end,
wu read out in Parliament by Sir Robert Peel in 1826, when he was attempting the fint
phase of the achievement of Bacon's propamme. See Woris, xiii, 57-7 1 •
 Immature and Ineffective Fields ofInlJuiry 377
details. Nor do they go through the same complex processes of
evolution and standardization as scientific results, and so the social
phase of their development does not involve such rigorous tests.
Also, they express a private understanding as much as a public
knowledge, and so their very statement will include terms with
important connotations of meaning which are lost when they are
removed from the context of their first announcement. Because of
this, they can be seriously misleading, when taken over by later
generations in ignorance of their history.
The other sorts of disciplines cannot hope to achieve knowledge
of the same sort as the matured mathematical and experimental
sciences; an aphorism will never have the permanence and objec-
tivity of a law. Whether they are thereby rendered inferior is
another matter; if their problems are important, and can be effec-
tively solved to yield some sort of genuine knowledge, then they are
worth pursuing. The same considerations hold for immature disci-
plines; although the results will be weak, and it is impossible to
predict when and how they will reach maturity, if the problems are
real then one continues the work as an investment for the future. The
question, then, is how to make this investment most effectively. I
have argued that it is unrealistic to proceed as if a discipline is
mature when in fact it is not, and that it is ill-advised to force it along
a path of mechanical imitation of matured disciplines. Whatever
model is adopted, the forcing of inquiry into the mould of 'research',
investigating small-scale problems totally within the world of intel-
lectually constructed classes of things and events, is unlikely to bear
scientific fruit, either in the present or for the future. By pretending
to be what it is not, the immature or ineffective discipline condemns
itself to remain in that state, rather than engaging in the sort of work
that would open paths to its achieving maturity of a character
appropriate to itself.

The Pretence ofMaturity

If scientific inquiry could be conducted in a social vacuum, the
failure to recognize immaturity, and to adopt methods of inquiry
appropriate to the condition, would be a sign of incompetence or
dishonesty in the leaders of an immature field. But the present
social institutions of science, and of learning in general, impose such
constraints that the growth and even the survival of an immature
 378 Science in the Modem World
field would be endangered by the simple honesty of public announ-
cement of its condition. 13 For these institutions were developed
around mature or rapidly maturing fields in the nineteenth century.
Their operation presupposes that good and first-class research is
possible, that it is governed by a set of stable and appropriate
methods, and that it has already produced a body of facts and
techniques in which students can be trained and examined. If the
represenmtives of a discipline announce that they do not fit in with
such a system, they can be simply excluded from it, to the benefit of
their competitors for the perennially limited resources. The field
would be relegated to amateur status, and thereby pushed over to the
very margin of the world of learning; it would be deprived of funds
and prestige, its members would need to pursue their inquiries in
the time left over from an unrelated job, and recruits would come in
only through self-education. It is true that the study of the natural
world as a whole was in just such a social condition in most places
up to the present century; but it could rely on men of independent
means as recruits, and there was no established institutional structure
devoted to the advancement of learning, from which it was being
excluded; the Academies included natural science in their concerns,
and the universities were at best teaching establishments. Hence
under present conditions it is natural, and tactically appropriate, for
the leaders of immature fields to conceal their condition, from them-
selves as well as from their audiences. We will later see that the
simple pretence of maturity is not the key to the institutional success
of an immature field; for this it must be believed in as a folk-science
by some educated public. But to secure such a belief, it must affiliate
to the general folk-science of 'Science', and this is equivalent to the
concealment of immaturity.14
IJ In the academic hierarchy, unpopular immature disciplines are in a social position
comparable to that of the category 'nigger': they must be gold to pass for silver. This
became clear to me when I learned of the dehOerations of a Committee of Professors on
the filling ofa vacant Chair in the History ofScience at a distinguished foreign university.
An acceptable candidate was required to have been an accomplished scientist flfIIl a
classical scholar; to be at home with historians of an desaiptiODS flfIIl also to be offering
a message relevant to modern problems; as well as having eminence in a special field of
raearch in the subject. One wise old man commented, 'They want a sheep with five
legs.' At the same time, powerful Professors on the committee were filling old and DeW
0Iairs in their departments with men who made no pretence of being anything but
productive raearch workers.
14 Some leading men in the field of psychology in Britain have taken the plunge into
public honesty. See G. Westby, BelunJiour Theories flU the Stfltus of Psytiolov,
IDaugural Lecture (Univcrsity of Wales Press, Cardiff, 1963). He concludes with a recom-
 Immature and Ineffective Fields ofInlJuiry 379
The pretence of maturity throws up a host of practical problems,
which aggravate the already severe difficulties of scientific inquiry in
such conditions. Recruits are generally given no warning that their
research work is likely to be very hazardous, and after some years of
producing results which inexplicably fail to consolidate into facts,
they can become demoralized; the real state of their field becomes
a shameful secret. Nor do they have the security of knowing that
their years of specialized training in the subject gives them a mono-
poly of practice in it, as in a learned profession or a matured science.
For the founders of the field, on whose insights all subsequent work
is directly based (without the enrichment that occurs in matured
fields), are likely to have been philosophers, polymaths, or amateurs,
rather than certified experts; and even at the present it is possible for
an amateur to crash into the field, or to analyse the practical
problems of its concern with more success than its practitioners. IS
Moreover, since the social mechanisms for quality control and
direction in the field cannot function properly, the safeguards against
the abuses of prestige are weakened, and the assessments of a lay
audience, based on popularizations, can be of more practical impor-
tance in the politics of the field, than those of the community of
experts. Under such conditions, the pretence of maintaining the

mcndation for 'patience and boldness' for this 'infant science', together with a link with
common sense experience. (I am indebted to Miss Helen Gower for this reference.) For
a aitical discussion ofAmerican sociology, in which many of the points made here about
the character and problems of immature disciplines are argued at length and with
examples, see Sociology on Trilll, eds. M. Stein and A. Viddach (Prenti~HaII, 1963).
(I am indebted to Mrs. J. Wootton for this reference.)
It is noteworthy that economics, by contrast, does not have any significant debates over
its methods and their appropriateness. The Festschrift to Oskar Morgenstern, ES$IIYs ill
Mlltllmulticill EctJtUJmics, ed. M. Shubik (princeton University Press, 196'7) contains •
variety of papers on theoretical economics and games theory, but not one discussing the
critical analysis of 0" the ACcurMy of EctnUnlli& ObsertJllhtmS. The reason may be that if
Morgenstern's recommendations were to be accepted, and economists did somehow
learn that a number is an estimate and not a fact, then both research and teaching would
need to be drastically overhauled, and many flourishing lines of research postponed for
at least • decade until worthwhile data could be obtained. But it cannot be said to be a
healthy situation when the published analysis and recommendations of an eminent
member of a discipline are recognized only informally, and ignored in reasearch and in
teaching.
15 In his History of Ezperimmt.1 Psychology, G. Boring identifies four very great men
in psychology's history: Darwin, Helmholtz, James, and Freud (p. 743). Only the last
two had any concern to be considered as 'psychologists', and Freud was, of course,
excluded from academic psychology all his life. Comparing Francis Galton to Wundt,
Boring said ofthe former 'He had the advantage ofcompetence without the limitation of
being an expen' (p. 462).
380 Science in the Modem World
social mechanisms appropriate to a matured discipline, and even
more the task of improving their real condition, are rendered yet
more difficult. 16

Teaching an Immature Discipline

It is possible to mask the effects of immaturity on the research
activities of a field; but in the work of teaching, which is necessary
both for the ttaining of recruits and for a vocational base, there are
problems which are both acute and insoluble. In some respects, the
tasks of teaching and of learning in an immature field can present
more challenge, excitement, and true education than in one where
there is an enormous body of standard information and tools to be
mastered before the student is accepted as competent to think for
himself. For in the absence of a great mass of established doctrine to
be imparted, teacher and student can participate in a common search
as near-equals, and the teacher's particular role can be Socratic
rather than magisterial. The student gains the experience of inquiry
into unsolved problems, and the teacher is continually refreshed and
rejuvenated by his dialogue with new students on new problems.
Unfortunately, there are so many obstacles to the realization of
this ideal state, that it can be achieved only in very exceptional
circumstances. Even for the attempt to be made, there must be a
prior public recognition that the field is so immature that there is no
large body of facts to be taught; and, as we have seen, such a recog-
nition involves such heavy costs for the field that only leaders who
are both courageous and powerful can afford to make it. Also, such a
style of teaching makes very heavy demands on the teachers; they
must have exceptional intelligence, commitment, and confidence for
the dialogue to be genuine and constructive. Discussions on the
materials of an immature science will, if they are free, range over all
levels, from data to methodology. Without skilful control, they can
easily degenerate into intellectual chaos at the one extreme, or be
structured into a set routine on the other. The teacher must be able
to steer this delicate path, year after year, avoiding the tempmtion to
use his superior knowledge of the teaching situation to guide the
discussion along familiar and relatively safe channels, and still pre-
16 In connection with his discussion ofacademic cliques in the social sciences, C. Wright
Mills quotes an anonymous comment about a nameIeIs statesman of social science: 'As
long u he lives, he'll be the IDOIt eminent man in his field; two weeks after he dies, no
one will remember him' (711 SDtiolo,;eM 1",.,;_ti"", p. log).
 Immature and Ineffective Fields ofIntjuiry 381
serving his personal stability. Under the conditions of our present
culture, it is even more difficult to find students who can participate
effectively and benefit from such an experience. They must be highly
intelligent, with minds well trained for abstract reasoning; otherwise
the teaching situation degenerates into an exchange of personal
feelings and impressions. But their previous training at secondary
school has been in the manipulative mastery of established doctrine;
an entirely new set of skills and attitudes are required here, and the
transition can be unsettling both intellectually and emotionally. It
requires unusual strength and understanding to master difficult
technical material which is patendy unsound, for the sake of the
training it gives in abstract thought, and as a corrective to naive
common sense. 17 Finally, even when the right teachers and the right
students come together, this is an intensely personal situation,
charged with excitement and with danger. It may succeed brilliantly
for one year or for five, but it is a most difficult thing to institution-
alize so that it can carry on for decades without becoming stale or
corrupted.
Even when it succeeds, this approach to teaching has its costs. As
an anomaly within an educational system conceived and organized
around the transmission of doctrine and the objective testing of the
mastery of information and skills, it is subject to a variety of external
pressures and internal strains. ·The problem of examination (an
essential part of the university's function as a certifying body) is
severe; for as the doctrinal component of the student's training
diminishes, the difficulties of objective and fair examination increase.
Also, the students are deprived of the security which comes from
mastering information and skills at a rate which can be tested against
a public standard. Their situation subjects them to all the anxieties
which afflict research students, and for which the institutional setting
of their undergraduate work provides no protection. A student for
17 See Ely Devons, 'Applied Economics-the Application ofWbat?', in Tile Lot" of
PtrSMUlI K,,01l'letlge (Michael Polanyi Festschrift) (Roudedge, London, 1961), PP.155-69.
Devons argues that 'what we are applying in theory are the elementary propositions and
common-sense maxims'; and finds these extremely diflicult to teach. 'It is only when the
student has been through a theoretical drill well beyond the elementary level that the
elementary notions really sink in deep.' And on the theory itself, he says earlier, '.•. but
the position is more pathetic when economists by using the language merely deceive
themselves. For the use of the language may give the illusion of a great understanding,
whereas in fact it often merely conceals ignorance in a mass of esoteric jargon. And it is
easy and dangerous to mistake description and classification of situations in a special
economic language as answers to problems' (pp. 164, 165, 161).
382 Science in the Modem World
whom such a course is inappropriate cannot settle down to the
achievement of a modest competence as in an orthodox under-
graduate subject; success is an all-or-nothing affair. Finally, until
the peculiar skills and attitudes gained by a successful student ofsuch
a course are recognized as being genuine and important, the graduates
will find themselves without a market for their labour. The strains of
preparing oneself for nothing are severe on all except those who are
confident of passing on to research, and those others who are not
concerned to fit in to society. In fact, undergraduate study in an
immature discipline is successful only at two extremes. For those
who merely want intellectual fun, without the grinding discipline
of an established subject, an easy course can be very enjoyable. And
for those who can quickly learn to grapple with unsolved and ill-
defined problems, a demanding course can be the best possible
education. The latter group form a natural elite, and are equipped to
function in modem society analogously to the social elite of former
generations whose education mainly consisted of training in the
personal and social characteristics ofleadership. But we are as yet far
from recognizing this quality; and in any event the difficulties of
distinguishing objectively between weak and demanding courses in
an immature discipline further delays such recognition.
In the great majority of cases, the teaching of an immature or
ineffective discipline will be attempted imitation of that of matured
disciplines. In the interests both of academic respectability, and of
the needs of the less gifted students, instruction will be of straight-
forward doctrine, of which students will be expected to achieve
manipulative mastery before voicing their opinions on open ques-
tions. But in immature fields, the straightforward doctrine does not
come to be as the outcome ofa natural evolution. There is no corpus
of basic standardized facts, utterly reliable in themselves, which can
be presented in a digestible form. Nor can the obscurities at the
foundations of the discipline be buried out of everyone's sight by a
mass of technical material. Those students who need to believe what
is taught to them in order to understand it are thereby indoctrinated,
perhaps for life. Those others whose critical sense impels them to
question the doctrine, find themselves trying to playa different game
from the one accepted by their colleagues and demanded by their
teachers. Should they persist, then crises, intellectual, emotional,
and social, are bound to result. Nor is it easy for a teacher, at least
for one who is not a naive believer. He must purvey the doctrine as if
 Immature and Ineffective Fields ofInlJuiry 383
it is physics, and attempt to answer the unanswerable questions of
the critical handful of students; and to the extent that he is aware of
the true state of affairs, the contradictions in his position will be
severe.
Technical and Practical Problems and their Hazards
It is when immature sciences are enlisted for the solution of
practical or technical problems that the most severe strains arise.
For this engagement inevitably leads to deceptions, of self and
others, compared to which those on the academic scene are but
minor slips. A discipline which is unable to establish facts even
within its closed world of controlled experience is much less capable
of genuinely drawing conclusions about the problems of a raw and
unstable reality. Of course, a master of the field may possess a
personal wisdom of its phenomena which enables him to perceive
the real situation and its problems more deeply than a person with no
special experience; but his conclusions are derived more from his
intuitive knowledge, built up informally over a long period, than
from any programme of piecemeal research. 18 Yet there are many
temptations to pretend, and to believe, that a Jarge-scale research
programme is necessary before adequate decisions can be taken on
an urgent practical or technical problem. Both scientists and
decision-makers have such temptations; and these, with their
effects, are worth reviewing.
First, a scientific field, like any other institution, has an innate
tendency to expansion, if nothing else than as a means of survival.
To refuse support for an expanded research effort, on the grounds
that the intended function of the research might not be performed,
is to possess such an otherworldly purity of character as would
prevent a man from reaching the point of choice in the first place.
Mixed with ambition is the ever-present hope that the field is really
on the point of maturation, and that engagement on such a problem
might bring it over the top. Even if the leaders of the field have
private doubts about its effectiveness in this respect, they can legiti-
mately reassure themselves that in the absence of its intervention
18 A striking example of the eruption of common-sense is cited by Wright Mills, in
Tile Sotiolo,;cal /""',;""t;01I, using comments by A. W. Gouldner on the work of the
'Grand Theorist' Talcott Parsons. It appears that when Parsons analysed post-war
Germany with. view to policy recommendations (or its social reconstruction, he used a
straight Marxist analysis of class Sb'Ueture, with no evidence of the 'normative structure'
10 strongly developed in his purely theoretical writiDp (pp. 43-4).
384 Science i" the Modem World
ignorance, prejudice, and self-interest would dominate any decision
to be taken. Moreover, the problem-situation will frequendy be one
whose solution will involve a large project, with the creation of a
corps of technicians, practitioners and experts. To have a share in
their ttaining will provide the field with influence and a firm
economic base;-and it will also protect the clients from the disastrous
effects of the agents' untutored common sense in the relevant
respects. And even if some of the answers are known already, they
williegitimarely carry more weight ifthey are derived as conclusions
from a mass of rigorously controlled data, than if they are offered as
aphorisms distilled from personal experience.
Those who must take the decisions on practical problems will
welcome research, for their own reasons. If nothing else, it imposes a
delay on the decision; and in the meantime the problem may dry up
and blow away, or they themselves may shift to a less exposed
position. Also, the conclusions of a research project, in which the
categories are tidily defined and recommendations neady listed,
seem more amenable to translation into the terms of a technical
project, and the articulation of routine tasks, than the aphoristic
generalizations of an informal study. And, to be sure, if other things
were equal (although in the case of immature disciplines they are
not) a scientific argument provides a more secure foundation for its
conclusions than an impressionistic essay.
Engagement on such technical and practical problems is certainly
necessary for the rapid institutional growth of a field, and under
favourable circumstances may contribute to its maturing. But it also
presents dangers, for which an appropriate term is 'hypertrophy':19
a rate of growth so rapid that the existing social mechanisms of
direction and quality control cannot perform their functions. In
response to the urgent calls for helpful research, a clever mediocrity
can build an empire and attain power and prestige at the expense of
those with more caution or scruples. In the absence of effective
controls from within the discipline, or from neighbouring subjects,
the worse excesses of entrepreneurial and shoddy science can occur;
and if the problems are partly military in significance, some of this
can be dirty as well.:&o The growth in volume ofresearch provides an
19 I am indebted to Mr. M. ]. Wilson, of Imperial College, London, for this tenD.
20 During the period of intense examination by students of Government-cponsored
raearch at the Massachusetts Institute of Technology (1g68-g), certain projects in
·political science' were judged by students to belong to aU these categories simultaneously.
 Immature and Ineffective Fields ofIntjuiry 385
opportunity for an expansion of the institutional apparatus, inclu-
ding an academic base. There soon appears a structure of post-
graduate and then undergraduate courses, mainly vacuous in content
and largely taught by a mixture of mediocrities, philosophers
manques, and entrepreneurs. All the contradictions inherent in the
teaching ofimmature sciences are made more acute, as the small core
of successful craft techniques and aphoristic wisdom is embedded in
a doctrine imitating a matured science, theoretical and applied.
'Pseudo-science' is not the most appropriate term for such fields, for
they are not essentially misconceived in the problems and objects of
inquiry. But if we describe them as 'cliche-sciences' we will charac-
terize their distinguishing feature: the genuine insights at their base,
which may well be valuable in the education of students whose
previous experience is utterly foreign to the area of inquiry, become
reduced to cliches as teachers and researchers in the field rub them
together in an attempt to produce a plausible fascimile of
scientific arguments. And their conclusions too, to the extent
that they are not vacuous academic jargon, will be nothing but
rearrangements of the cliches that constitute the materials of the
field, organized for the best performance of the political functions
of a result.
The graduates of courses in such cliche-sciences then emerge as
manpower-units with spurious qualifications for miring their places
as technicians, practitioners, or experts on the growing industry of
vacuous research or misconceived technical problems. ZI In such
circumstances we can speak of corruption; for there is a sufficient
penumbra of uncertainty about the nature of the enterprise, that
while it is not universally recognized as a straightforward racket,
there is an awareness of solnething false about it which is best not
discussed too openly.
The hazards of hypertrophy of an immature field are thus very
real, especially in the present period when new practical problems,
21 An indication of these dangers has been given by Paul LazarsCe1d: 'But sociology is
not yet in the stage where it can provide a safe basis for social engineering•••• It took the
natural sciences about 350 years between Ga1i1eo and the beginning of the industrial
revolution before they had a major effect upon the history of the world. Empirical social
research has a history of three or four decades. If we expect from it quick solutions to the
world's greatest problems, if we demand of it nothing but immediately practical results,
we will just conupt its natural course.' The quotation is taken from C. Wright Mills,
The So&iolo,;tlll /",.';utUm, p. 100, with a reference back to 'What is Sociology', a
duplicated paper at Oslo University, 1948. MiUs italicises 'a safe basis for social engi-
neering', as the relevant passage in the context of his discussion.
386 Science in the Modem World
and new technical problems concerned with human behaviour, are
recognized at an increasing rate. The dangers extend to the whole
world of science, for a rise to dominance of immature disciplines
could, regardless of the intentions of their leaders, wreck the subde
system of direction and quality control at the highest level, whereby
the health and vitality of the entire scientific community are main-
tained. And like any genuine practical problem, this one admits ofno
easy solution. It is not merely that any new practical or technical
problem requires some new scientific materials to aid in its solutions,
and that these will tend to come from new and hence immature
fields. Another factor in the situation is the intimate relation of
practical problems and immature disciplines with folk-sciences;
and we must discuss these in some detail in order to analyse the
problem.

Poli-Science
The category of,folk-science' has long been recognized; but it has
always been believed to apply only to preliterate or subliterate
cultures. It is a part of a general world-view, or ideology, which is
given special articulation so that it may provide comfort and reas-
surance in the face of the crucial uncertainties of the world of
experience. In earlier ages, some ofthe leading sciences and educated
arts grew out of popular folk-sciences; among these were the ancient
art of prediction, as well as other forms of magic, along with the very
sophisticated arts of alchemy and astrology. Mer the relegation of
such fields to the status of pseudo-sciences in the seventeenth
century, it appeared to scholars that there was an absolute distinction
in character between rational, progressive 'science', and the super-
stitious and retrograde folk-sciences. It could not be denied that in
the historical order genuine science emerged out of folk-science;
but the causes of this deep ttansformation remained a mystery, and
survivals of the old in the new presented a serious embarrassment.
But the functions performed by folk-sciences are necessary so long
as the human condition exists; and it can be argued that the 'new
philosophy' of the seventeenth century, with its disenchanted and
dehumanized world of nature and its appreciation of closely con-
trolled experience, itself functioned as a folk-science for its audience
at the time. For, as it appeared then, it promised a solution to all
problems, metaphysical and theological as well as natural; and it
 Immature and Ineffective Fields ofIntJUiry 387
gained a stable social base long before its achievements were com-
mensurate with any of its claims. 22
Indeed, we may say that the basic folk-science of the educated
sections of the advanced societies is 'Science' itse~ in various senses
derived from the seventeenth-century revolution in philosophy.
This is quite explicit in figures of the Enlightenment such as
Condorcet; 23 and a basic faith in the methods and results of the
successful natural sciences, as the means to the solution of the
deepest practical problems, can explain the absence of genuinely
critical studies of the natural sciences as cultural and social pheno-
mena until very recendy. 24 Experiences which are inexplicable
within the categories of this 'Science' (as those indicating powers of
mind independent of matter) are denied significance or even exis-
tence, for they constitute a challenge to an entire world-view, and
hence threaten chaos and disintegration. A more visible and vulgar
effect of the role of 'Science' as the dominant folk-science of our
time is the status that the tide itself confers on any field of
inquiry or practice. Grabbing for such smtus, with its attendant
social and material benefits, is most widespread in America,
where the folk-science of 'Science' seems strongest and where
control on tides is weakest; 'mortuary science' is a well-known
case of a craft dubbed science, with its attendant academic
apparatus. 25
sa J. Ravetz, 'What was "The Scientific Revolution" ?', ltUlUI" Jowul ofHill., of
Sdmee, 1 (1g66), 15-21.
23 See Condorcet, Espisse tl'un TllbleQu historipe tles /Wo"is tle l'esprit 11"";" (1795),
ed. o. H. Prior (Boivin, Paris, 1938); 'Neuvieme ipoque', pp. 145-202. 'Toutes les
erreurs en politique, en morale, ont pour base des erreurs philosophiques, qui elles-
memes sont li&s l des erreurs physiques. · .' (p. 191).
24 It is significant that through the entire nineteenth century none of the great
sociologists analysed science as a social phenomenon. Comte took his idea of 'positive'
from the successes of the natural sciences; and for Max Weber natural science wu the
purest realization of 'rationality'. The earliest attempt I know at analysing the style and
metaphysics of natural science as the product of • particular culture was that of Max
Scheler, Die Wissmsformm fItUl tlieGesells,IuIji, 1926. See J. R. Staude, Mu S,heler,'"
ItJlelle,tul Portrllit (The Free Press, New York; Collier MacMillan, London, 1967),
pp. 18cH)1. Scheler's insights seem to have derived from his long-standing contempt for
the bourgeois and his belated conversion to a mystical philosophy. There may have been
a current of Marxist analyses already; shortly afterwards quite similar ideas were
expounded by Marxist writers.
25 A list of American 'sciences' has been provided by C. Trusedell: 'A short nndom
search of the catalogues of Graduate Schools delivered, in addition to the ubiquitous
"Social Science", "Political Science," and "Computer Science," some special examples:
"Meat and Animal Science" (Wisconsin), "Administrative Sciences" (Yale), "Speech
Sciences" (Purdue), "Libnry Sciences" (Indiana), "Forest Science" (Harvard), "Dairy
 388 Science in the Modem World
The relations between disciplined inquiry and the folk-science
aspects of a subject are naturally complex and variable. 26 To make
the distinction clear, we can consider the case where the same field
could, in different aspects provide both the technical setting for a
scientist's research, and also a folk-science for his personal belief.
This was quite common in the period when science was close to 'the
philosophy of nature' in spirit and name. The quotations from
Helmholtz and Einstein given earlier (pp. 39 and 66) each show
how reflection on some more or less extended part of science, either
in its methods or results or both, could provide comfort and reas-
surance. Indeed, the Helmholtz quotation, and our discussion of
ethics in science, show that without some personal adherence to an
ideal of science as providing some very important knowledge, this
very demanding work could not be sustained at the highest level for
more than a few generations.
The external groups for which a subject functions as folk-science
can vary enormously in their size, sophistication and influence; the
particular folk-science itself can be more or less central to their
ideology, and more or less durable. The style of work in the folk-
science will depend on such circumstances; for a sub-literate
audience it needs fewer trappings of academic jargon and titles,
while a sophisticated audience requires a reasonable facsimile of a
leading branch of'Science', such as physics. But in any case, the clue
to a subject's being a folk-science is that belief in it is quite indepen-
dent of its achievements in producing facts or knowledge, or in
accomplishing genuine solutions of technical or practical problems.
Since the functions performed by a folk-science are so different from
those of the solutions of problems investigated in a disciplined style,
the relevant criteria of adequacy and value of its results will be cor-
respondingly different. Value is determined by the degree to which
a problem-situation is central to the experience of the audience; and
adequacy by the success in offering reassurance and the promise of
understanding. For an educated audience, the mere existence of its
folk-science as a recognized academic discipline may be sufficient to
these ends.
Science" (D1inois), "Mortuary Science" (Minnesota). EsSilYs in the History ofMechanics
(Springer Verlag, New York, 1968), p. 75.
26 A brief but illuminating analysis of the character of the interaction between the
'scholarly' and 'folk' aspects of a discipline, from • Marxist point of view, applied to
Marxism i~ is given by A. Gramsci, 'Marxism and Modern Culture', in his Tire
Motlml Prinee .rul otller "";ti,,,s (Lawrence & Wishart, London, 1957).
 Immature and Ineffective Fieltls ofI"'l-Y 389
On occasion, adoption as a folk-science by some influential group
can provide a maturing field with social support (in the way of
prestige, jobs, finance, and recruits) sufficient to enable its firm
establishment; such seems to have been the case with history in the
nineteenth century. Conversely, some particular results ofa matured
field may (with suitable interpretation and vulgarization) suddenly
become a component ofa leading folk-science, and embark on a new
career independendy of the thoughts and desires of their creator;
such was the case with 'Darwinism'. 27 But there can also be conflict,
especially when a challenge to a scientific orthodoxy is made from a
base in a popular folk-science, and uses a style of inquiry and tactics
of struggle appropriate to that base. The ensuing struggles can be
very bitter, for the threat is not merely to aschool, but to the autonomy
of the world of established science in its goals and methods. Even
when a challenge to scientific orthodoxy is made by a man thought
to be a scholarly crank attacking a strong field from a shallow
popular base, the shadow of a threat can be sufficient to cause a
violent reaction; this was seen in the extreme measures taken by the
American community of physical scientists against Emmanuel
Velikovsky and his theories of recent cosmological catastrophes. 28
Immature sciences are even more vulnerable, of course; and (as we
shall see) more closely related to folk-sciences. Hence in them the
sttuggles between competing schools can involve not only doctrines
and personalities, but rival ideologies, their related folk-sciences,
and perhaps their political associations as well.
Folie-Science and Ideological Conflicts
The conflicts between academic science and strongly based folk-
sciences provide many insights into the social situation of science,
27 The interpretation of a chapter in intellectual history in termI of the folk-lCience
functions of an idea has been done by J. w.. Burrow, Ew/lltitnJ tUUl StJtiet,: • SIfIIl, of
VitltwUl" Soeial Tuor;, (Cambridge University Press, 1966). He states, in the Preface,
'It will be argued in this book that the seeds of modem sociological theory were to •
ccmsidenble extent implicit in the doctrines which were becoming current in the 18601,
but they were stifled by the overriding needs of an evolutionism which provided what
the Victorians sought in theories of society' (p. xiii).
On 'Social Darwinism' itse~ the standard work is R. Hofstadter, S«W D"";";"" i8
Ameri,." Thought (Harvard University Press, 1944): for the English scene, see G.
Himmelfarb, D",,,,;,, .utu D",,,,;,,;_ Revollll;tm (OJatto It Windus, London, 1957),
especially cbs. 14 and 19.
28 For a well-argued presentation ofVe1ikovsky's side ofthe story, see Tile YeliitJ'Dsi,
Ajffli" ed. Alfred de Grazia (University Boob Inc., New York; Sedgwick It Jackson
Ltd., London, 1966).
390 Seim&e;" the Modem World
and into some insoluble problems inherent in that situation. Such
conflicts have been little studied by historians from this point of
view, and so are worth some extended discussion here. These
conflicts take two very different forms. The first occurs when the
results of disciplined scientific inquiry contradict the beliefs of a
folk-science, usually a popular one which is also adopted by the
established cultural organs of society. The second sort occurs when
a popular folk-science, perhaps with a radical social and political
message, attacks academic science for its style and connections as
much as for its content. The former sort of conflict is better known;
the history of such conflicts can be read as a series of ttiumphs for
intellectual integrity and freedom ofthought, against popular super-
stition and imposed dogma. Indeed, there was a strong tradition of
such histories, starting in the Enlightenment and continuing through
the nineteenth century; for the sttuggles were real and bitter, and the
camp of'science' stood for the highest values ofcivilized life. 29 How-
ever, the work of science, when it is so engaged, is something more
and something less than that ofa self-contained 'positive' discipline.
There is an added dimension of social and philosophical involve-
ment; and those who inherit the successes of that sttuggle, both in
the knowledge achieved and the secure social position won, cannot
honestly wear the mande of the heroic reformers of the past. On the
other hand, 'positive' science can justly claim autonomy from society
and the state, on the grounds that its particular results are neutral
in their social and ideological effects, while the maintenance
of the activity of research is beneficial in a general and diffuse
way.
If we are to understand the moral problems involved in the
sttuggle for 'freedom in science', we must recognize that those who
attack established or official folk-sciences, are advancing doctrines
that are 'ideologically sensitive': their effects may be disruptive of
the stability of society. They may claim the freedom to make their
attacks, but they must then base their claim on some higher principle.
One such principle is the absolute right ofanyone to proclaim Truth;
another is the ultimate value to society resulting from the free clash
of conflicting doctrines. But to claim the right to utter unsettling
doctrines, while denying any responsiblilty for their effects, comes
perilously close to the arrogation of power without responsibility, a

w".,
at The dusic in this type of historical Hterature is A. D. White, A Hi"., of tile
ofSMt&e .,;" T1Ieoltv i8 C1IrisIerulo&
 Immature and Ineffective Fieltls ofIntpUry 391
ttaditional variety of immorality.30 For some generations up to the
present time, most academic natural science has been 'ideologically
neutral'; and indeed this neuttality has been invoked by propagan-
dists for science, as a reason for its complete autonomy of goa1s from
outside interference. The exceptions to this neutrality have been
little studied; one which may be historically significant is the close
association, in the early twentieth ·century, of the science of genetics
with a generally reactionary and elitist folk-science of 'eugenics'.31

30 Galileo early received clear warning of the ideological sensitivity of his Copernican
doctrines from his friend and protector, Maffeo Barberini, later Pope Urban VIII. In
March 1615, Galileo received a letter from Ciampoli reporting a mnversation with
Barberini, advising him not to 'exceed the limits of physics and mathematics', for all
author's assertions on matters in the Scriptures can, on publication, be transformed out
ofrecognition. He cites the example ofGalileo's analogy between the earth and the moon
in the casting of shadows on the surface: 'somebody expands on this, and says that you
place human inhabitants on the moon; the next fellow starts to dispute how these can be
descended from Adam, or how they can have mme off' Noah's ark, and many other
extravagances you never dreamed of.' See S. Drake, Distovtrits.tUl OjMlitnu ofGlIlile.
(Doubleday, New York, 1957), p. 158. There is no evidence that Galileo remgnizcd the
force of this problem up to the time of publication of the DWope tnJ tile TIIHJ Wtwill
Systems. There is, however, an odd passage in·the later DistOflses tnJ TIIHJ N", Stimtts,
which may indicate that after the struggle was over, Galileo finally saw what it was all
about. It mmes when one of the interlocuton praises Salviati (Galileo's mouthpiece) for
a particularly elegant demonstration, and remarks how little such achievements are
appreciated. Salviati replies with some reflections on those who claim to be competent
in a field of study, and yet still publicly deny new truths in spite of beHeving them at
heart, 'merely for the purpose of lowering the esteem in which certain others are held by
the unthinking aowd'. (Edizione Nazionale, p. 204; Dover·ed., p. 169.) This particular
audience is hardly ever mentioned in Galileo's published works; and so one is tempted
to suppose that at that very late stage he had finally discovered its relevance to the
struggle which he had lost. One may put it that he had remgnizcd the category of 'dirty
truth'. Modern students mme to a study of Galileo affair with healthy libertarian
prejudices; for the education ofthose in my tutorials, I put the question, 'Would you try
to prevent someone from giving away mloured mmic:s depicting scenes of sexual
violence against children, at a junior school?' Those who would are then in a position to
appreciate that there was a genuine tragedy in Galileo's conflict with the Church.
By mntrast, Descartes saw the point of ideological sensitivity as soon as he heard the
news of the condemnation of Galileo. He prompdy suppressed his completed work
Le Mtmtlt; and three years later, when explaining this, mentioned 'certain persons to
whom I defer, and whose authority governs my action no less than my reason governs
my thoughts ..• '. It seems clear that he had never previously imagined that there could
be a contradiction between these two governing principles, especially for someone so
shrewd, cautious and honourable as himseU: The aucial text here is the opening section
of the sixth part of the DiseOfls de 111 Mlt1uHJe, in which Descartes finds himse1farguiDg
both sides of a question, without a clear aDSwer coming out of the application of his
Method.
31 Of this particular association between science and social affairs, which was con-
veniendy forgotten during the denUDCiations of Lysenko's campaign against orthodox
genetics, some indications are given by Paul Gary Wenky, 'Nlllw, and PoHtics between
392 Seim&e ;" the Modem World
Even to this day, the social sciences are ideologically sensitive, and
affixing the label 'positive' to anyone of them does not alter this
reality. Also, the problems of 'reckless science' and of the sciences
associated with runaway technology have resurrected old moral
problems, and seem likely to raise ideological issues analogous to
those involved in earlier conflicts with popular or official folk-
sciences.
When academic science finds itself on the defensive against
challenges made by popular folk-sciences (as distinct from counter-
attacks by the servants ofan official orthodoxy), it becomes involved
in the deepest problems of politics; and these show how precarious
is the neutrality and universality of scientific inquiry and scientific
knowledge. For academic science has hitherto inevitably been res-
bicted to an elite leisured class and its occasional recruits; and its
conventions and etiquette are meaningless to the less privileged
sections of society. The institutions of science have usually been so
far removed from the perceptions of the lower orders that both sides
are happily ignorant of its position in a class-divided society. But
there have been occasions when a section of the prevailing academic
science has been attacked, for its content, its style, and its social
position, by representatives of a radical movement demanding its
replacement by a populist folk-science. The Paracelsian tradition in
early modem Europe provided a basis for such challenges, of which
the best known is that of the 'sectarians' during the English Civil
War. 31. A similar pattern can be discerned in the attack on the
Acadbnie des Scim&es during the French Revolution;33 and again in
the career of the Soviet agronomist Lysenko. In each case the chal-
lenging folk-science related to a 'romantic' philosophy of nature,
combining a stress on crafunan's manipulation, a personal involve-
ment in the work, a democracy of participation, and a disttust of
the Wan', N""t, U4 (1g6g), 4M-72. I am indebted to him (or this reference; he bas
work in preparation which will show the connectioD fully.
Ja For the English Civil War, the semiDaI paper is P. M. RattaDsi, 'Paracelsus and the
Puritan Revolution', AmlJiz, I I (1963), 24-]2. The theme is developed further by the
lime author in 'Politics and Natunl Philosophy in Civil-War England', Anes ~u Xl-
C"""is l",trUritnuU tl'HiSl,;,e des Stinltes, ii (0sI0Iineum, Warsaw, 196'7), 162-6. The
basic printed clocumentl (or this CODttoYeny are the polemical worb on the universities,
publisbed in the early 16501; see A. G. Debus, Stinlte IIfIIl EtlwMitnl i8 ,lie Stw1JIemlll
C""",: ,lie Wei",,-W.4 DeNte (Oldboume, London, 1970).
JJ For the ideological coo8icII over ICieDce in the ndicaI phase ofthe French Revolu-
tion, lee C. C. Gillispie, 'The E'"1e1tJliilU and the Jacobin Philosophy o( Science: I
Study in Ideas and Consequences', in CritieIJI Pr,i1nlu i8 ,lie Hi".:! ,f Stinlte, eel.
 Immature and Ineffective Fields ofInquiry 393
abstract or mathematical reasoning. Viewed from the point of
34

view of academic science, such movements are all doomed to end in

tragic failure, since they are incapable of producing worthwhile
scientific knowledge. 35 But it may be that their leaders have been
less concerned to achieve the same sort ofspecialized knowledge in a
different style than to bring about a total revolution in human life,
and make their contribution as Utopians rather than as reformers.
In the folk-history of academic science (that is, the folk-science of
its own past), the defence ofthe autonomy ofscience has been seen in
terms of attacks by simple obscurantists, or by the State. This other,
more troubling sort of conflict, in which the enemies of science rise
from a popular base and seek to bring back the philanthropic ideal of
knowledge, has been little explored. But the involvement of contem-
porary science, not merely with elite culture, but also with the
technical and administrative government of a nation, is bound to
M. Clagett (University of Wisconsin Press, 1959), pp. 255-89. Gillispie shows that
Lavoisier's reform of chemical nomenclature was seen as a move involving politics and
ideology: denying that chemistry is a branch of 'natural history', accessible to every man;
and converting it into a branch of abstract, quantitative, elitist physics. Independent
mnfirmation of this effect of the reformed nomenclature is provided by • letter of
Thomas Henry, the Manchester manufacturer, to James Watt, of 1790: 'Chemistry
being a science so intimately connected with many ofthe Arts, and consequendy studied
by many illiterate men, should have itllanguage plain and intelligible and not made up by
words mmpounded from a dead language which none but men of Ieaming can under-
stand.' Q!1oted from A. E. Musson, and E. Robinson, Sdnlee IIfIIl TeelmtJlov jn llu
ltlllrutrilli Rt'VOllllitnJ (Manchester University Press), p. 242.
34 For Lysenko, see Z. A. Medvedev, Tiu Rise "till F"O ofT. D. LyseUo (Columbia
University Press, 1969). The 'romantic' aspect ofLysenko's style does not appear in his
main polemical writings, but it was detected by C. A. Waddington. In a discussion ofthe
cluster-planting of trees, Lysenko is quoted by him as baving made the point, '••• that it
would be against the great laws of orpnic life if members of the same species competed
with one another, and did not help one another, and he went on that nothing is worthy of
being called ttue science unless it exhibits the great underlying order of the cosmos in
general'. Waddington felt that Lysenko's underlying philosophy was not so much
dialic:tical materialism as baving 'a pretty strong Savour daived from onhodox Russian
theology, with God left out'. See C. H. Waddington, 'Talking to Russian Biologists',
Tlu ustmer, 6g (1963), 119-21. The French biologist Marcel Prenant aIIo discussed
this topic with Lysenko, and came away with the same impression, comparing Lysenko
with Bernardin de Saint Pierre (see Medvedev, ope cit., p. 168). It is ~blc that German
N"'Wpllilosopllie, nther than religion, was the prime source of Lysenko's biologic:&!
philosophy; it came into nineteenth-century Russia with political radicalism, and if •
link could be established through agronomy, • strong circumllantial cae could be
argued. The matter is a delicate one, for Marxism, like othernineteenth-century German
ntionalist philosophies, concealed its connections with N"",pllilosopllie ancl a6:ially
despised it.
3S The futility of the 'romantic' style ofscience is the theme of a lengthy sun ey nf the
history of scientific ideas, by C. C. Gillispie, Tile Etl,e of Olljenivil1 (PriDc:ctoD UPi,....
sit)' Press, 1960).
394 Seimte;" the Modem World
lead to political challenges to its position. One of the main targets of
the Cultural Revolution in China was the prevailing style ofacademic
teaching and research; it was argued that these were perpetuating
the rule of a bureaucratic elite, who would inevitably lead the nation
along the 'capitalist road'.36 In the event, no rival folk-sciences made
a strong appearance, and some branches ofWestern-type technology,
as the nuclear weapons programme, were protected by the State; and
so the Cultural Revolution did not proceed through its full course
in this field.
The Chinese experience offers some clues to the patterns of future
political challenges to academic science. As the traditional handicraft
industries are destroyed by technological change, the folk-sciences
based on them, which run continuously back to the magic of earlier
times will inevitably vanish. Also, since a modern society, unlike most
earlier ones, is so crucially dependent on a sophisticated physical and
social technology, an attack on its institutional and cultural basis can
succeed only at the price of destroying the fabric of social life. But
these conditions in themselves do not prevent the appearance of
revolutionaries who do not know of, or care about, such consequen-
ces. With the rapid growth of an educated and relatively leisured
class in the advanced societies, radical movements can fasten on to
folk-sciences which relate to man's spiritual existence in quite
sophisticated terms. The first example of this is the succession of
movements and groupings, strongest in America, which has pro-
duced 'beats' and 'hippies', not all of whose members can still be
J6 It is iDsufficiently realized, by Marxist u well u by liberal scientists, that the very
real 'democracy' ofscience is Itrictly bourgeois democncy: an abstract equality for thole
who have bad the opportunities enabq them to use it. I can argue this with a paradox:
• boy DOW growing up in an urban black ghetto bas a greater chance of becoming
President-of the UDited States of America than o( becoming • cfistinsuished physicist.
For it is easy to imagiDe • brilliant youth teaehiDg himse1f to read during adolescence and
then embarking on a sua:eaful poIitic:al career; but (or physics he would be well past
his most aative years by the time he was in. position to do original research. It may be
that this contndictioo in the Iocial position of science bas prevented historiaDs sympa-
thetic to 'popu6It' teDdeDc:icI in the study of uature from achieving a fully synthetic
account o( them. This can be seen in J. Needham, Stinlte IIfIIl CiviNs.,;. i8 emu,
Y01. ii (Cambridp UDiftl'lity Press, 1956). He gives a muterly acaKIDt of the different
IIpectI of Taoism, iDcIuding European panUeIs to 'empiricism and mysticism' and the
ideals of Iocial benefit; but since their political attaeb were on the Confucian ,oliti&"l
philC*)J)hy, he can usimilate the democracy ofmodem science to that conceived by them
(pp. 86-.132 ).
For an account of the QUnae Cultural Revolution in relation to science, see C. H. G.
<>ldbam, 'CUaa Today: Sc:icDce' (Cantor Ledure),Jowul oftill R01M Sodny ofArts
116 {196!), 666-h.
 Immature and Ineffective Fields ofI1IIJ-Y 395
youths. Although their message is one of gendeness and love, they
are far more deeply radical than the hyper-militant activists, who
complain about the system of intellectual culture only because it is
not tailored to their desires. 3 ?
Up to the present, however, the important folk-sciences of the
educated classes have not been involved in such fundamental
sttuggles. Sometimes a fully matured academic science could also
function as a folk-science; thus there was a great stream of 'popular
science' of the eighteenth and nineteenth centuries, in which the
wonders of physical science were cited either to show the cleverness
and beneficence of the Creator, or alternatively His nonexistence,
depending on the audience. And when sections of the educated
classes (perhaps in association with some of the self-educated) took
up an 'unrecognized' science as a folk-science, as phrenology, it was
not used as a lever for the displacement of the whole system of
official culture.
An intermediate situation, of great importance, is the adoption of
an immature discipline as a folk-science. This is indeed necessary, if
it is to survive beyond the personal endeavours of its founder and a
small band of enthusiasts. Once the discipline is established, its
'disciplined' and 'folk' aspects maintain a subde but deep interaction.
The adherents of the folk-science will describe and hence perceive
the world in terms conditioned by the conclusions of the science;
their common sense is thus pardy a product (at some removes) of
disciplined inquiry, if not scientific then at least philosophical. On
the other hand, the objects of inquiry of the discipline were them-
selves likely to have been distilled from the common sense of the
founders of the field, who could well have had an ideological moti-
vation for their work; and so the folk-science as it existed at the
inception of the field can have a continuing influence. Moreover, in
an immature field the objects of inquiry are not modified and
enriched solely by the internal processes of the solution of problems
and the creation of facts; the internal weakness of the field makes it
more subject to external influences, even in the modification of its
objects of inquiry by the changing currents of common sense in its
lay audience, as conditioned by their personal experience and the
teachings of popular expositors. Thus the 'folk' aspect of a science
can react back on its 'disciplined' aspect; and although individual
37 For a penetrating and sympathetic account of this movement, see T. ROSDk, Tile
MUi", of" Ctnmler Cullllt'e (Doubleday, New York, 1969; Faber, London, 1970).
396 S,im&e;" the Modem World
productions in the separate aspects will superficially have a very
different style and content, they are to a great extent merely per-
forming different functions within a single enterprise.
The human and social sciences show this admixture of 'folk' and
'disciplined' aspects to a strong degree; and this is significant in
view of their being basic folk-sciences for large and important
sections of the educated population. The question of the nature of
the groups adhering to them is a complex one; but the 'swing' from
(natural) 'science' among adolescents, and the disproportionate
support for student revolts among those in the social sciences, are
clear indications of a strong nucleus. Their influence is far more
pervasive than is indicated by the university scene alone. For some
decades, psychology has been an important folk-science in America,
generating an enormous handbook literature on all aspects of the art
of living. Economics is doubtless the folk-science of all those com-
mitted to an economy planned to any degree; in spite of the vacuity,
or irrelevance of most of its theory, and the patent unreliability of its
statistical information, it ranks as the queen of the sciences in the
formation of national policy.38 Debates over economics as a disci-
pline are almost all debates over political economy; and while the
critics, of the far Right and libertarian Left, ruthlessly expose its
fallacies and inadequacies, academic students are carefully ushered
through the accepted doctrines in blissful ignorance of these
difficulties.
The ideological function of a subject as folk-science can vary
enormously; in the nineteenth century economics was the 'gloomy
science' which demonstrated the necessity of the harsh conditions of
JI See Ely Devons, 'Statistics as I Basis for Policy', Lloyds BaM RtvieJl' Guly 1954);
reprinted in EUilYs jn EetnUlflli&s (Allen It Unwin, London, 1961), pp. 122-37. We read:
'Considered in this light there seem to be striking similarities between the role of
economic statistics in our society and some of the functions which magic and divination
play in primitive society'; and Devons argues the analogy at some length, along the lines
which I describe as the functions offolk-science. His criticism was ofeconomic statistics
as they are aetua1ly used; in principle they might be used more 'scientifically'.
One of the leading and constant dogmas of applied eamomics in Britain since the war
is that wages should be kept down, because ofthe obvious and necessary relation between
wages, internal prices, and export prices. Any rise in unit wage costs in manufacturing
will, we have been told incessandy, lead to such increases in export prices as to cause loss
ofmarkets, national bankruptcy, etc. A set of charts of these three variables in the Lloyds
BtMi RI'DinI of April 1964, p. S4t showed that over the previous decade they had
correlated reasonably well for the United Kingdom; but for Italy, France, and Japan
they went their own ways independendy. As befits pnctitioners of an immature folk-
1Cicnce, the eamomic advisers to the State took not the slightest notice of this incon-
venient information.
 Immature and Ineffective Fields ofI1Uj1Ury 397
unrestricted free enterprise, while now it is the folk-science of
planning (although, to be sure, many distinguished economists are
quite conservative). Similarly, although the great tradition of nin~
teenth-century social philosophy was generally conservative in cast,
in the present century sociology has become the folk-science of
liberal reformers. 39 Thus when the American Supreme Court
needed a doctrine on which to base its decision that racially segre-
gated schools cannot be 'equal', and had available no inherited
common sense capable of being translated into constitutional prece-
dent, it boldly produced an explicidy sociological argument. Al-
though such a procedure is more honest than the traditional one of
twisting precedents to suit one's prejudices, it also carries with it the
danger of endangering the autonomy of the law, and making it sub-
servient, in explicit detail as well as in principle, to the conclusions ofa
folk-science, which, as a scientific discipline, is still immature. Such
support at the very highest levels ofsociety gives a practical influence
to the favoured disciplines, out of all proportion to their scientific
strength. Yet the situation is not one for which blame can or should
be assigned; for the close relation of practical problems, immature
disciplines and folk-sciences make such practices inevitable.

Immature Fields and Prfl&ti&al Problems: Some Haurds

We can now return to our discussion of the difficulties that arise
in the attempt to 'apply' sciences that are ineffective or immature, in
the solution of practical problems. This application takes place first
through programmes of research when the problems are first
publicly recognized, and then in the training of experts for the
technical projects through which it is hoped that the problem will be
resolved. It is only natural that the disciplines involved in new
practical problems should· be immature; for the novelty of the
problem takes it outside the domain of competence of established
fields of inquiry.
The difficulties that ensue are only partly the result of the con-
dition of immaturity or ineffectiveness of the field that assumes the
load of research and indocttination. For the recognition of a new
practical problem is a political phenomenon, involving a change in
the real world as perceived by an important section of the public.
I' See R. A. N'JSbett, Tile Sorio1tJlitM T,UiIitnt (Baic Boob, New yon. 1«)66;
Heinemann, London, 1967).
398 S,ientt in the Modem World
Indeed, it frequendy occurs that a problem-situation is recognized
and analysed by isolated scholars for years before it suddenly comes
to exist as a political event. The eventual popular recognition may
be a result of serious disturbances of the accepted order of things
(such as mass action in the form ofriots, an economic crisis, or a war);
or a particularly eloquent account of the problem may reach the
public just as its previous cares are losing their dominance on its
perceptions. However, even for that new perception to take place, it
must be coherent with the informal categories of the ideology of that
public; and it is perceived in the terms ofits accompanying explana-
tion, cast in the relevant folk-science. The function of such explana-
tions is to minimize the threat displayed by the problem-situation;
this is done either by explaining it away altogether, or, if this is
impossible, to give assurance of the possibility of its comfortable
resolution. If these tasks cannot be accomplished, then the ideology
of that public faces collapse, and the people themselves are threat-
ened with the disintegration of their world.
Thus the emergence of a new problem-situation is not merely a
political event; it also presents new problems for the folk-science of
those concerned. Simi1arly, the solution of the practical problem is
twofold: both to identify and remove the causes of the distressing
situation, and to restore the threatened sense of security. One aspect
is handled by the scientific inquiry leading to the establishment of a
practical project; and the other by the folk-science. But these two
aspects intermingle; the announcement of the investigation of the
practical problem is itself reassuring; and since the objects of
inquiry ofthe immature field are already so closely related to those of
the folk-science, there is a tendency for any inquiry to be conducted
along those lines which support the underlying ideology of the
threatened group. Also, there are many situations where the beliefs
and expectations of the individuals of the relevant public influence
the reality they collectively experience; and so the reassurance
offered by the folk-science may have a chance of being a 'self-
fulfilling prophecy'. In cases where this applies, the function of the
relevant scientific disciplines in the solution of practical problems
cannot be entirely different from those of the ancient 'pseudo-
sciences' where the client's belief (or credulity) was essential to
success. Should the academic study of the discipline ignore these
similarities, it can encounter pitfalls whereby a new 'science' of
supposedly rational man becomes nearly isomorphic in its sttueture
 Immature and Ineffeetive Fields ofIntpUry 399
to an ancient pseudo-science of man involved in a humanized and
enchanted world. 40
When an immature field takes on the task of expanding its
research effort for the solution of some urgent practical problem,
there will be a tendency for the outcome of its labours to be a weighty
argument establishing the conclusions that its sponsors and its
public wanted all along. 41 This result seems to be independent of
the style of the inquiry, whether it is large-scale sponsored research
by experts on the American pattern, or the smaller, largely amateur
inquiry by committees of mandarins on the British pattern. The
British style does at least avoid the worst dangers of hypertrophy;
and it has the additional benefit that the leaders of the inquiry, free
of any professional involvement in its conclusions, can continue to
learn about the problem from the debate that follows the publication
of their report. 42 However, the tendency to confirmation of the folk-
40 Some years ago I argued that nuclear strategy is isomorphic in ill sttueture to
classical asttology; see J. R. Ravetz, 'The New Asttology', SSRS NeJlJslmer (Society for
Social Responsibility in Science), No. 140 (Aprillg64). I now believe that this identi-
fication is mistaken, and unfair to asttology. Ptolemy's defence of the science in his
Introduction to the TnrfllJiblios (ttans1ated by F. E. Robbins; Loeb Classical LibJUy,
Harvard University Press; Heinemann, London, 1940), pp. 3-35, can apply to any human
science whose predictive power is as yet limited; and in the mathematical elabontion on
common sense foundatioDS, classical asttology provides a model for any more recent
mathematical social science. The ancient science to which nuclear sttategy is truly the
successor is gematria, as applied to conflict situatioM. There, a number could be
assigned, by a standard coding procedure, to the name of each contestant; and there was
a prediction algorithm based on comparison of the two numbers. If both sides used the
same code and predictive system, the science would tend to confirm its predictions, and
also avoid unnecessary bloodshed. See the Seeret... SeeretOlVlJ of Roger Bacon, eel.
R. Steele (Oxford, 1920), p. 251. At the time of this writing, nuclear sttategy bas lost all
credibility and respect; and the nuclear arms race is recognized as depending only on ill
own insane momentum. Those who did not live through the later 19501 may find it hard
to believe that this patendy absurd pseudo-science was taken seriously in councils of
state; but it WIS, and its subsequ~t discrediting is a useful precedent for those who wish
to criticize such new pseudo-sciences and c1iche-scienc:es as may arise.
41 In his review of Ayres, TetlmtJlogie.l ForetllSli"g IIfIIllmlr-R.",e pr-.;tJ8, M.
Shubik (op. cit.) puts 'Project Hindsight' (a study of the origins of technology in 'pure'
research) into perspective. 'In general all one needs to know about a review project is who
is the sponsor, who is inventing the criteria for judgement, and who is running the
project; at that point the odds are overwhelming that the conclusions can be guessed
before the study had been done. It is my belief that Project Hindsight has many of the
signs of a hatchet job and Ayres is being naive in his treatment of it.'
42 Of course, only a minority will take this opportunity. Among these is Dr. F. S.
Dainton, the chairman of the committee which produced the report on the 'swing from
science' in England, E.'1 into the Flo1lJ ofC.rulitltltes i" Stimee .tIIl TetlmtJloD mt,
Higller E~lItllritm, Cmnd. 3541 (H.M.S.O., 1968). The report itself made a number of
recommendations for changing the seamdary-school syllabus, with the implication that
these would be sufficient to arrest the 'swing'. After an inteDSive public discussion ofthe
400 Seim&e in the Modem World
science is not absolute, except of course where the folk-science itself
is tied to a rigidly enforced political orthodoxy. Independent
inquiries can reach conttary conclusions, and offer them to the
public. But these are relatively easy to ignore, since their categories
and language will be foreign to the original audience. Only if these
inquiries r~late to the folk-science of some other group will they
force themselves into recognition. But their challenge will be per-
ceived mainly as a political one; and in the ensuing debate, the
different levels of argument, from the scholarly to the popular and
political, will be inextricably entangled and confused.
It is only to be expected that the application ofscientific inquiry to
new practical problems should be even more hazardous than the
management of deeply novel results within science itself. To the
extent that the investigation of problems loses its protective frame-
work of accepted and successful methods, it becomes exposed to
pitfalls of every sort. On being associated with an influential folk-
science, an immature field, in chaos internally, experiences the
additional strains of hypertrophy, and its leaders and practitioners
are exposed to the temptations of being accepted as consultants and
experts for the rapid solution of urgent practical problems. The
field can soon become identical in outward appearance to an estab-
lished physical technology, but in reality be a gigantic confidence-
game, combining the worst features of entrepreneurial and shoddy
science. The dangers of such conuption are at present more acute
for some of the social sciences and technologies (especially those
using mathematical and computational tools) than for the natural
sciences, since they are related to the most urgent practical problems
and they lack a base in fully matured disciplines. Protection against
these dangers is not to be found in mechanical imitation of the
methods of the matured disciplines; pseudo-research is one of the
symptoms ofthe diseased state, and sophisticated criteria ofadequacy
of results (as 'fa1sifiability of theories') are irrelevant to situations
where conflicting purposes and ideologies are central to the prob-
lems, and the discipline's function as folk-science cannot be elimi-
nated. To thread one's way through these pitfalls, making a genuine
contribution both to scientific knowledge and to the welfare of
society, requires a combination of knowledge and understanding in
report, Dr. Dainton came to lee that the problem is much deeper, and ascribes the
'nriD( to the DeW mood of the JOUDIt who lee ICicDce both u irrelevant to human
COIlCeI'DI and also u hard work. See SMIte JOIInIIIl (October 1969), p. 13.
 Immature and Ineffective Fields ofI""wry 401

so many different areas of experience, that its only correct tide is

wisdom.

Conclusion
From our analysis, it appears that the condition of ineffectiveness
or immaturity in a field is not a simple absence of some desired
characteristic of a matured science. Indeed, a field can be kept in
such a state by inappropriate methods, adopted partly to achieve a
state imagined to be that of a mathematical-experimental natural
science, and partly to pretend that it is already there. Under such
circumstances the very real difficulties of achieving worthwhile
results, of teaching, and of operating the social mechanisms for
the guidance and control of research, are made still more severe.
However, the external constraints on such a field, in the present
social arrangements of science, would make it very difficult for the
leaders ofan immature field to adopt methods appropriate to its con-
dition, recalling the old distinction between 'philosophy', 'history',
and 'art'. When immature and ineffective sciences are enlisted in the
effort to resolve practical problems, their difficulties are still further
compounded. In addition to the hazards of hypertrophy and conup-
tion, an immature academic field is particularly liable to the com-
plex influences from its associated folk-science; and to become
submerged in cross-currents of political and ideological conflicts.
The picture is a gloomy one, but since immature and ineffective
fields are due to be involved in public affairs to an increasing extent
as our social problems become ever more complex, an awareness of
their limitations is necessary if their application is to systematically
produce more good than harm.
Also, with an appreciation of the naturalness of the state of
immaturity, we can see it as one phase ofa full cycle of development,
rather than imagining that every science emerges from a futile pre-
history, onto a permanent plateau of the 'positive' state. With the
state of immaturity corresponding to the infancy of a field, from
which only the hardy survive, we can see the phase of maturity
giving way, in its tum, to one of senescence. The lasting achieve-
ments of the phase of immaturity will be a few aphoristic insights of
a philosophical character, and perhaps some tools and techniques;
those of the phase of maturity will be a mass of positive knowledge.
But when the basic insights are exhausted and the leading problems
shift elsewhere, the field enters senescence, useful only for its
402 SGimte;" the Modem WorllJ
standardized information and tools purveyed through teachers. 43
This cycle is easily observed on the small scale, in descendant-lattices
ofproblems and their associated schools; but it may also apply, on the
large scale to whole sections of the world of scientific and scholarly
enquiry. To be involved in a field just entering maturity is the most
rewarding career for a scientist; for then one can make great achieve-
ments at relatively little risk. But estimating the points of transition
between phases is a very delicate task; a field or area ofscience which
is approaching senescence is a dreary place; and immature fields
with the hope of imminent maturation are, with all their attendant
ha2ards, the place where the greatest challenge is found.
4J The senescence of academic physics has been indicated by the distinguished
physic:iat, A. B. Pippard: 'But we should not forget ~t the great en of academic
physics may well be drawing to its close, and that we are very likely entering a new en to
be domin'ted by fundamental advances in molecular biology, biochemistry and pharma-
coIou, all of CDOI'IDOUI industrial potentiality.' The last phrase indicates that the
eenesceDCe may extend to all of academic science, although the context of the remarks is
the problem of scientists in industry. See A. B. Pippard, 'Innovation in Physics-based
1Ddustry', Alma D of'Tbe Swum Report', pp. 10C>-3. I am pUeful to Dr. J. Wootton
f. this refereDc:e.
 Part V
CONCLUSION: THE FUTURE OF
SCIENCE
 CONCLUSION: THE FUTURE OF
SCIENCE

FOR several centuries, the understanding of science has been con-

ditioned by a belief in the separateness of knowledge and society.
The faith in the attainment of human knowledge which is absolute
and unconditioned inspired the pioneers of the new philosophy out
of which natural science grew; and it has been the easiest line of
defence of the autonomy of science against its many enemies. That
simple faith is no longer adequate for its function of maintaining the
integrity and vitality of science. Attempts to refine it through purely
epistemological analyses do not provide a basis for defending science
against the dangers and abuses arising from its new social conditions;
but to consider science as merely a special branch of industrial
production would lead to its speedy degeneration. The argument of
this book has attempted to exhibit the ways in which genuine
scientific knowledge can be a product of a social endeavour, and yet
embody ttuth, at least within the fundamental metaphysical frame-
work of the civilization in which it is achieved. From this analysiS
we have been able to study the conditions under which science
can advance towards knowledge, to identify diseases and abuses to
which science is subject, and to examine the special features of the
application of science to the solution of technical and practical
proble~.

Recapitulation
Our analysis necessarily started with an absttaetion from a com-
plex reality: considering the investigation of a problem as the unit-
task of scientific inquiry. We saw that this work comprises several
distinct phases, each involving sophisticated craft skills. Contact
with the external world is made in the production of data; but this
must be converted into information, and then used as evidence in
an argument. The argument concerns artificial objects, intellectually
consttueted classes of things and events; and it is about these that
14-&U.P.
406 Ctmt/usion: The Futur, ofS,ien,e
the conclusion is drawn. Since no argument in science can be
formally valid, and hence no conclusion necessarily ttue, the
acceptance of conclusions must be governed by criteria of adequacy.
These impose a complex and subde sttueture on the argument, and
ensure the avoidance ofthe known pitfalls which can be encountered
in manipulation or in inference. They belong to the body of craft
knowledge of the methods of the field, along with particular tech-
niques of using tools and with other conttolling judgements such as
those of value. The methods are informal and even tacit, and are
transmitted interpersonally rather than publicly; they are incapable
ofbeing tested scientifically themselves, but arise out ofthe collective
craft experience of the field.
The conclusion of an adequately solved problem is still far from
being knowledge or even a fact. The research report on a solved
problem must be assessed by a referee before it is certified through
publication in a recognized journal. Even then, it must prove its
significance (by being put to use), its stability under testing and
repetition, and its invariance under the changes in conceptual
objects which inevitably occur as new problems are investigated. Of
all the facts which are so established, the great majority remain
within the descendant-lattice of problems deriving from their
original, and sink into oblivion when that field is exhausted and
forgotten. Those facts which survive to become scientific knowledge
have a different path of evolution. Rather like successful tools, they
are also extended to other fields, in standardized versions performing
a variety of functions and taking diverse forms. When such facts
have survived the demise of their original problem and its descend-
ants, and remain alive through their many uses, they are recogniz-
able as knowledge. It is paradoxical that the different extant versions
will be incapable of being reduced to a single, standard statement;
and that the obscurities latent in the original formulation will
frequendy remain unresolved. It is also paradoxical that the whole
process of evolution and selection is accomplished by fallible in-
dividuals, and governed by a craft knowledge of methods. Such a
conception of the nature and origins of genuine knowledge runs
counter to the hitherto dominant ttaditions of the philosophy of
science and epistemology. But in them, the basic problem was how
an individual could quickly achieve ttuth or the best substitute.
Here, the guiding principle is 'veritas temporis filla'; and as the
daughter of time, transformed and tested by a great variety of
 Co",lus;on: The Future of S,ien" 407
contacts with the external world, recognizable scientific knowledge
emerges from a complex and lengthy social endeavour.
For simplicity this first analysis was restticted to matured fields
of 'pure' scientific inquiry, and it presupposed the presence of
social mechanisms whereby the private purposes of individuals
would be harmonized with the collective goals of the endeavour. An
examination of these mechanisms was necessary as a preliminary to
any analysis of the conditions under which the health and vitality
of science can be maintained. We saw that in the protection of the
intellectual property embodied in an authenticated research report,
the inherited formal system of journals and citations must be
operated by an informal etiquette if it is not to be abUStd and
destroyed; and the inttoduction of new forms of property lacking
the controls of the journals requires a most refined etiquette if they
are not to lead to a degeneration of the work. The management of
novelty has in the past presented some of the most severe practical
problems for science. We saw that neither the assimilation of old
materials to new, nor the choice between competing research
strategies, can be accomplished on the basis of general rules; but
that destruction and conflict are inevitable in science, and the social
organization and style of work in each particular area will determine
whether the outcome of a 'revolution' will be renewed vitality or
stagnation. The social task of quality conttol in science bears most
directly on ethics, for the immediate private purposes of most
individual agents are served by skimping, however slightly, on the
quality of their accomplished tasks. No formal system of imposed
penalties and rewards will guarantee the maintenance of quality, for
the tasks ofscientific inquiry are generally too subtle to be so crudely
assessed; nor will the advantages to an individual of a good reputa-
tion ofhis group be sufficient to induce a self-interested individual to
make sacrifices to maintain it. Only the identification with his col-
leagues, and the pride in his work, both requiring good morale, will
ensure good work. Science possesses a hierarchy of quality control,
informal except at the lowest level where research reports are
assessed for publication; and the conttollers are conttolled by
rewards of prestige in various ways. At the top of the pyramid of
conttol are the leading scientists at the leading universities, who
conttol each other and their fields by the most informal oftechniques.
They are neither answerable to their inferiors, nor sttictly account-
able to those who provide their support; and their work of direction
408 Conclusion: The Futur, ofScience
and control requires both wisdom and the highest ethical com-
mitment if it is not to degenerate. The ideology and social context of
this ethical commitment are inherited from an earlier age; and the
conditions of industrialized science present them with problems
and temptations for which their inherited 'scientific ethic' is totally
inadequate.
Science becomes directly involved with society at large when it is
applied to the solution of technical problems, involving the pro-
duction of the means for the performance of a function, or practical
problems, involving the achievement of the purposes of individuals
or groups of people. Each of these other sorts of problems have their
characteristic cycles of investigation, their appropriate criteria of
quality, and their particular pitfalls. The most sophisticated technical
problems, in which the scientific component is sttongest, are en-
countered in those industries where the automatic mechanisms of
quality conttol through a competitive market are weakest; in them
there is the danger, not merely of runaway technology advanced
without regard for human welfare, but the corruption of the activity
of technical problem-solving itself. The investigation of practical
problems, and their solution through large-scale practical projects,
encounters every pitfall of scientific and technical problems, and
then some peculiar to itself. Conflicting ideologies and purposes are
at the heart ofevery urgent practical problem; they lack the accepted
criteria of quality for their solution; the sciences involved in them
are usually immature; and in their execution they are prone to
distortion by the natural tendencies of bureaucratic operation.
Because of the increasing recognition of new practical problems,
immature sciences are assigned tasks which they are not strong
enough to accomplish properly; to their internal difficulties (ag-
gravated by the necessary pretence of maturity) are then added those
ofhypertrophy. When involved in the solution ofpractical problems,
they also function as folk-sciences; and the resulting confusion of
the different sorts of problems and their appropriate styles of work
can result in total demoralization and corruption. These difficulties
and dangers are directly relevant to the future health of the ttadi-
tional, established, matured natural sciences. For their internal
difficulties of recruitment and morale, as well as those of their
relations with society at large, are practical problems of the sort
analysed here. Their resolution is urgent, because of the delicacy
and wlnerability of scientific inquiry; but the pitfalls to be en-
 Conclusion: The Future of Scien" 409
countered here are no less dangerous than on any other practical
problem.

Science in History
Our analysis of genuine scientific knowledge showed that it is the
outcome of a lengthy process operating through history; and we
have also indicated various ways in which scientific inquiry as a
whole is an historical phenomenon: c;onditioned by its social and
cultural environment, and subject to cycles of growth and decline.
The very long period of the flourishing of the matured natural
sciences was plausible evidence for the comforting beliefthat science,
in its academic and positive period, had truly reached evolution's
end, and would thenceforth experience a simple progress onwards
and upwards indefinitely. The 'idea of progress' with which the
rise of modern science was intimately associated received its mortal
blow in the First World War and its aftermath; but in science
itself the assumption survived for nearly another half-century.
Appreciating that that long 'golden age' of science is now definitely
ended, we can see it as one temporary phase in the history of man's
attempts at understanding and control of the perceptible world
around himself. Extending back to remote antiquity, this history has
the common pattern of continuity and change, and gains and losses,
as its successive phases appear. Seeing ourselves in a new phase of
this history, we can face its problems as inherent in its conditions,
and not as merely accidental difficulties to be removed by exhorta-
tions or by administrative devices.
While recognizing the novelty of the problems of industrialized
science, we would be quite mistaken to think that the whole social
activity of science has been completely ttansformed in the last two
decades. A large part of scientific research proceeds as before, in a
social context which is still mainly 'academic' rather than industria-
lized. The radical difference is that certain new tendencies resulting
from industrialization are developing rapidly, and that the self-
consciousness of science, as reflected in the pronouncements of its
leaders, has changed from a simple optimism to a ttoubled uncer-
tainty. Even if these developments continue and intensify, one
cannot predict with any assurance just how serious their effects will
be within any given time. The work of scientific inquiry is now
embedded in a very large social institution which performs other
410 Conclusion: The Future ofScience
essential functions in an advanced society, including higher
education and the investigation of technical and practical problems.
Even ifall these sectors encounter increasingly serious problems they
will continue to receive support from society so long as they are
considered as performing these functions better than any feasible
alternatives. Historical change can take a long time to work itself
out; and we must avoid the naive rationalism characteristic of
radical reformers, who believe that once the insoluble problems of an
institution have been exposed it will soon pass away. We might recall
that the Catholic Church was conscious of a deep internal crisis as
early as the later twelfth century; the incompetence and corruption
of the clergy had already led to powerful movements of reform
and schism. Yet several centuries were to pass before 'Reformation'
achieved a permanent, independent base; and the Church
survived that, to continue as a powerful force up to the present.
Of course, the Church had coherence, wealth, and power in a
way that the social institution of science does not; but if we
are assessing the prospects for science over a period of a few
decades to come, we must keep in mind the enormous inertia of
any established institution in all but the most revolutionary of
contexts.
We must also remember that the world of science is a very varie-
gated one. Some fields are capital-intensive, and so very vulnerable to
the effects ofindustrialization; while others can produce outstanding
work with small investments for each project. Again, some are
closely related to technical problems, while others can proceed in
peaceful uselessness. National styles in science, which were very
marked even among the successful nations before the complete
domination of academic science, may emerge again so sttongly as
to condition the sort of work which is successfully done in different
places. It is well known that the greatest sttength of America lies in
technical problems: both the development of physical devices and
the organization of work and management. But American science is
particularly prone to the dangers of enttepreneurial, shoddy, and
dirty science. By conttast, British science is sufficiendy small and
poor (in comparison) to be led by an institution (the Royal Society)
still resembling a club; and in this context the problems of con-
ttaction, and accommodation to industry, may be managed with
more finesse. Again, in the Soviet Union the political pressures on
intellectuals are so crude, that the leaders of science have a natural
 Ctmtlusion: The Future ofScience 411
social and political role as spokesmen for an Enlightenment in
classic eighteenth-century terms. I
The history of science provides yet another caution against over-
simple predictions of the effects on science of the changes in its
context. While it is relatively easy to give plausible explanations of
the gross features of scientific activity in terms of its social and
cultural environment, this becomes progressively more difficult as
one tries to account for work of lasting quality, and the productions
of genius. We have mentioned earlier that some of the greatest
scientific work of all time was conducted within a metaphysical
framework which would now be rejected as superstitious and anti-
scientific. Yet even those who were searching for the divine har-
monies of the celestial motions, or for the material location of the
world-soul (as Kepler and Gilbert, for example) could, by talent and
discipline, achieve results which became incorporated into the body
of genuine scientific knowledge. Similarly, although men of ability
will generally do better work when they are part of a vigorous com-
munity, enjoying prestige and leisure for their researches, some ofthe
greatest advances have come from men working under difficult con-
ditions in nearly complete isolation: such men as Copernicus,
Mendel, Galois, and Lobachewski are cases in point. Hence, even
if the goals of 'positive science' are totally displaced in scientific
inquiry, and the major communities of science experience crises of
finance and morale, there may yet be scientific achievements which
will last for centuries to come. However, there is no known means of
encouraging genius through adversity; and it would be dangerous in
the extreme to conclude that the quality of scientific work would
improve if scientists were left to starve in garrets.
Science in Society: the Problems ofMorality
With all due caution in the face of the complexities of historical
change, we can proceed to indicate the deepest problems that
science will face as the process of industrialization develops, and to
speculate on how they might be resolved. For this we will first need
to analyse the inherent tensions in the relations between scientific
inquiry and the society at large which supports it.
• See A. Vucinich, 'Science and MonJity, a Soviet Dilemma', Snmee, 159 (15 March
c:urren.t movements for autonomy of science, in
1968), 1208-12, for a discussion of the
the context of the Russian traditions. For an example of this style of work, see A. D.
Sakharov, Progress, CtNzistmee "rul ["tellen"," Free__ (Norton, New York; Deutsch,
London, 1968; Penguin, London, 1969).
412 Conclusion: The Future ofScience
Ofitself, scientific inquiry is not a self-sustaining social institution;
neither wealth nor po,ver are derived directly from the activity. It
requires support from society at large, or at least some wealthy and
powerful section of it, if it is to exist. From lay supporters, then,
science requires first of all 'legitimization', if its practitioners are not
to be relegated to a despised or abhorred fringe ofsociety. More than
mere tolerance is required; resources must be invested in scientific
work, both in providing paid research time for the practitioners, and
in supplying their specialized equipment. Finally, an increasing flow
ofrecruits, drawn from the adolescent population, is necessary if the
work is not to stagnate and die. In exchange, science can offer the
promise of assistance in the solution of technical and practical pro-
blems. This may be direct, or indirect; thus high-level teaching,
which in recent generations has been considered to depend on an
association with research, contributes to technique; and the con-
tribution of science to national prestige, or to the strengthening of
the nation's official ideology, helps in the solution of practical
problems. But we notice that these return offerings of science to
society are not, and cannot be, dominant components of the general
goals of the work, and still less of particular research projects. Some
of them may well be present, in varying degrees, in the work of
particular individuals or schools, especially in the conditions of
immaturity and in the endeavours of a genius. But should a large,
established field, depending on the efforts of many research workers,
allow its criteria of value (and hence of adequacy as well) to be
dominated by such external functions, the work which results will
not be science. It may have excellence of a different sort, or it may
be quite corrupt, depending on conditions; but it will contribute to
the advancement of knowledge only very incidentally.
Thus the social position of science is really quite precarious.
Scientists are not professionals ofthe ttaditional sort, who can justify
their position by the serving of the purposes of clients; nor can
science be conducted on a large scale in a social context analogous to
that of the fine arts, providing prestige to particular pattons. Science
not merely requires very tangible support in return for quite intang-
ible returns; but the different components of its support will in
general derive from different sections of society. Each of these will
need to be furnished with propaganda appropriate to its tastes; and
this internal complexity, together with the great variety of social con-
texts within which science has operated, have produced a great
 Conclusion: The Future ofScience 413
variety of themes in the literature of justification and defence of
science. The dominant themes from earlier times relate to the
ideological functions (and dysfunctions) of science or of natural
philosophy; for until the middle of the nineteenth century only a
very few fields (as chemistry, itself only recently established as a
science rather than a craft) could make any plausible claim to con-
tribute to industrial production. In the later nineteenth century an
accommodatio' i with industry was recognized as necessary; and in
the present period there is a strange mixture of 'images of science'
purveyed to its different audiences.
In earlier times, the principal threats to the autonomy of science,
or rather natural philosophy, have occurred when some fields were
considered as ideologically sensitive, endangering the established
religion; and in this respect they were involved in the practical
problems of their age. Those with responsibility for the spiritual
welfare of the lay public would use all the means at their disposal to
contain or eliminate the dangerous doctrines and their perpetrators.
Such measures ,vould be more successful in places where the Church
had an established machinery for handling doctrinal crimes, and the
power to enforce its decisions; hence the Catholic Church has had
an unfair reputation of outstanding enmity to free inquiry. As a
result, there developed a belief in a close association between
scientific inquiry and independent, rational, or free thinking in
general. Propagandists for any of these ttaditions have assimilated
the martyrs of science (most notably, the very complex figure of
Galileo) to their cause; and some important traditions within the
folk-history of science have imagined the community of science as
necessarily composed of individuals who are selflessly and fearlessly
devoted to Truth, against Authority and Superstition.
This identification rests on several basic fallacies. Scientific inquiry
must have a subtle and complex assessment of the strength of
evidence deriving from accepted authority; and in this it is similar
to any other work where partly new problems are being solved. Only
in sectarian religion and in teaching can the work continue success-
fully for any length of time without encountering the problems of
the management of authority; and of course the total rejection of
authority, whereby every assertion must be examined as an equal
claim to truth, quickly yields chaos. Also, scientific inquiry is
ideologically sensitive only accidentally and occasionally. In England,
for example, the propaganda for science, purveyed from the
4 14 Conclusion: The Future ofS,;mee
seventeenth through the nineteenth centuries, argued that the
contemplation ofGod's creation could not but induce to ttue religion.
Of course, in England hardly anything gets really sensitive ideologi-
cally; and English divines could use the persecution of Galileo as
evidence for the wickedness of Rome. And when English scientists
were confronted with the practical problems of uncomfortable
theological implications of their work, they were far more likely to
devote their energies to a reconciliation than to make some specialized
results a fulaum on which they would move all heaven and
earth.
Towards the end of the nineteenth century, the applications of
science to technical problems were increasing in number and in
power. Spokesmen for science had the delicate task of securing ever-
increasing support from the community on the basis of such useful-
ness, while still preserving the autonomy of science itself. We recall
the brilliant speech of Helmholtz (quoted on p. 39 above), where he
reminded his audience of the apparent 'uselessness' of Galvani's
experiments on animal electricity, which yielded the eleCb"ic tele-
graph; but then warned them that the scientist himself must not be
expected to search for anything but new knowledge about nature.
This sincere plea for the social support of science, on the grounds
of its accidental social benefits, is characteristic of the period of
matured academic science.
With the advent of industrialization in science the claim of the
technical applicability ofscience in general did not need to be pressed
(although particular fields still need to justify their requests for
support in these terms). Although there have been continuing dis-
cussions about the proportion of resources which should be devoted
to 'pure' or 'basic' or 'undirected' research, there is a general
recognition among policy-makers of science that it performs a
variety of useful functions and so deserves some minor share of the
budget. However, another audience has suddenly become crucial for
the continued well-being and expansion of science: its potential
recruits. Among them, there is a significant fraction who see the
applications of science quite differendy from the nineteenth-century
optimists such as Helmholtz. Not the telegraph, but the Bomb, has
become the type-example ofthe leading technical problems in which
science is engaged. Hence a new, negative, and defensive theme has
been developed in the propaganda for science: its neutrality. Of
course, this will be purveyed to lay and juvenile audiences, in the
 Conclus;on: The Future ofScience 415
hope that they are unaware of the firmly realistic terms in which
'science policy' is now cast.
Since the claim of 'neutrality' is the last defence against recogni-
tion of the political and moral problems of science consequent upon
its industriali7~tion, we can expect it to be advanced for some time
to come; and it is worth closer analysis. The attempt to disclaim
moral responsibility for the effects of scientific work has been made,
not merely for 'pure' science, but even for work on technical
problems with a military function. This extension of the domain of
moral isolation of science is too implausible to survive. To be sure,
one can agree that in one sense 'it is the height of folly to blame the
weapon for the crime', Z if for nothing else than that inanimate things
are not appropriate objects of moral judgements. But those who are
engaged in making weapons in the knowledge that their main
function will be in the commission of a crime are subject to moral
judgements and sometimes to legal proceedings as well. If this were
not the case, then the defence of Adolf Eichmann, that he was
merely engaged on a technical problem oftransport, indifferent to the
fact that it was a one-way transport of Jews and others to the gas-
chambers, would be a valid one.
It is more plausible to assert the neuttality of science for that work
which is governed by purely internal criteria of value, so that in the
choice of problems the possible technical functions are either un-
known or irrelevant. Even here, there is an area of ambiguity; for a
particular research worker may choose to work on a problem for its
functions in the advancement of the field, and for its political
functions for his own career, while being aware that the investing
agency is interested in the problem for its possible technical func-
tions. If these technical functions are morally dubious or bad, can
he claim immunity? The ignorance of consequences is not always a
valid defence in law, and the mere absence of deliberate intent to
malefaction is an even weaker defence.
However, there are severe difficulties in the way of making precise
and fair moral judgements on scientists, individually or collectively.
If nothing else, our experience of these problems is extremely short,
and we possess neither general principles nor case-law for their
interpretation, whereby responsibility and blame can be assigned in
any but the most blatant cases of dirty science. Also, the division of
2 See Sir Peter Medawar, 'On the Effecting of An Things Possible', Presidential
Address to the British Association, 1969, NnIJ Seimtisl (4 September 1969), 465-7.
416 Conclusion: The Future ofScience
labour in large-scale technical problems is extremely fine; so that the
scientist who publishes a result generally has no more knowledge of
its possible functions than does a process-worker assembling a
standard component of a device. This does not mean that the posi-
tion of the agent is one of moral neutrality.; rather that it is morally
indeterminate. However, to the extent that his research is related
to technical applications, the area of indeterminacy decreases, and
the scientist's responsibility becomes defined. Once the scientist is
aware of the likely consequences of his work, his sole disclaimer of
responsibility can be along the lines of, 'I was only following orders.'
This is no longer likely to be acceptable as a defence, in science as
anywhere else.
These considerations apply strongly to scientists who are em-
ployees of a 'mission-oriented' research establishment; but for those
whose experience is in a community of science still enjoying an
academic style these moral problems are remote and philosophical
rather than immediate and personal. There, research problems
and personal achievements are evaluated by internal criteria, re-
gardless of the motives of those supporting the work; the hard work
and strict self-discipline are supported by a refined ethic; and results
are shared with colleagues independendy ofall the boundaries which
divide mankind. There, it seems, worthwhile work can be done,
insulated from the moral squalor of ordinary life.
But even to the extent that the moral neutrality of academic
science is real, it creates moral problems on its own; and the deep
connection of science with the culture in which it is embedded
involves science in its basic problems of justification and survival. The
practical irrelevance of most of the results of scientific inquiry is a
blessing to some, but an agony to others. What can be more selfish
than to tum one's back on the sufferings of humanity, devoting
one's talents to narrow tasks which will bring immediate rewards to
oneself and only the most remote and unknowable benefits to one's
fellow man? For a young scientist with a strong social conscience it
requires an extremely strong faith in the human value of scientific
knowledge to justify such a career. 3 Once this moral dilemma is
recognized, it can be alleviated by various sorts of good works, but
3 J. G. Crowther, Tile SodtllReliltiom ofStimte (MacMillan, London, 1941), has some
sharp words on this problem: 'Young scientists who abandon science for politics often
prove to be mentally unstable, and after a few years of bohemian agitation become
conspicuously conservative. Conduct and opinions that appear to be based purely on
moral sentiments are nearly all mspect' (p. 644).
 Conclusion: The Future ofScience 417
never completely resolved. Moreover, the remoteness of academic
science from human concerns is itself a result of its own traditions,
deriving from its particular niche in a particular society. The openness
and internationalism of science are admirable and valuable in them-
selves; but they are not the same thing as a sharing with all humanity.
Rather, they are methods of social behaviour of a small group
operating within European literate culture. Now that national
boundaries within that culture are of decreasing emotional signifi-
cance, the ttanscending of them is correspondingly less impressive.
And to the extent that this culture as a whole is subjected to moral
judgement, for its involvement in 'various sorts of colonial and class
oppression, and for its creation of a runaway technology, academic
science will be inescapably implicated as well.
It is quite likely that those of the present generation of elder
statesmen ofscience who invoke its 'neutrality' are trying to reassure
themselves as much as any audience of potential recruits. Con-
sciousness always lags behind reality, and its adjustments are usually
abrupt and painful. The conception of science as an essentially
academic enterprise, and of 'the scientist' as an academic researcher,
has persisted unchallenged in all the literature about science until
very recently, in spite of the ttaditional industrial connections of
several major fields (particularly chenlistry, but also physics and
biology), and in spite of the fact that the great majority of those who
have ever earned their living through their scientific skills, have done
so in technical work. 4 Even the interpenetration of science and
industry can be traced back to the later nineteenth century, and can
be seen as growing continuously since then. An awareness of a new
condition of science came only when a series of dramatic events,
such as nuclear weapons, and new technical problems, such as the
planning oflarge-scale scientific research, obttuded themselves. The
self-consciousness ofscience is still trying to cope with these changes;
the purely technical problems of decision and control are difficult
enough in themselves, and the deeper practical problems of re-
sponsibility and morality are only beginning to be grasped.
In the short run, the easiest response to such problems is to hope
4 N. D. Ellis, in The Snmtifi& Worltr, has shown that the Royal Institute of Chemistry
and the later Institutes of Physics and of Biology were created with the co-opention of
the major employers of scientists; and their ethical principles were framed for the
'professional employee'. The academic scientist was not their concern; and these
institutes are now subject to some strain because the majority of 'Q:S.E.'s' in industry
have no professional status, but are nther 'scientific workers'.
4 18 Conclusion: The Future ofScience
that they will go away, and in the meantime to try to get the best ofall
worlds. Such a course of action, which was almost certainly not the
result of a conscious policy, was taken by British university science
teachers in the post-war period. Still believing in science as a
genuinely liberal education, they expanded their departments with
State funds intended to provide more units of trained manpower, and
in practice taught Honours degree courses designed for that small
fraction who could proceed to research. The instability of such an
arrangement was revealed after only a decade or so; and whatever the
outcome of the pressures for its alteration, the world of science has
suffered no pub.lic discredit thereby. But the same sort ofconvenient
wishful thinking, applied to the understanding of science as a whole,
can have dangerous consequences, including a corruption of the
whole work.
The sort of corruption which can occur in science has little in
common, superficially, with that which is recognized in public life.
Hence it is necessary for us to analyse the concept briefly, to show
why it is relevant to the problem of science. s We can say that an
activity is corrupt when the actual goals ofthe tasks accomplished are
contrary to the professed social functions to a degree that a public
trust is beuayed. Corruption is occasionally flagrant, but more
commonly it exists in a penumbra ofambiguity; both the divergence
between the final causes, and the awareness of that divergence, are
ill~efined. Because of this, a man may work in a corrupt situation
without himself being corrupted. 6 If he is ignorant of the state of
5 The only worthwhile analytical study of corruption is The Autobiography ofutU:ol"
SteJfnu (Harcourt, Brace, New York, 1931). Steffens was able to gain experience of
corruption in American public life through his work as a journalist; but he could rise
above mere 'muckraking' because he had previously spent years of (informal) study and
thought on the problems of ethics. Although he never achieved a coherent solution to the
problem of corruption (which is an interesting problem because, as Steffens found, the
c:orrupton frequendy have more honesty and personal integrity than muckrakers and
reformers), his book is a mine ofexperiences and insights. My attempted formal defini-
tion derives from his interview with 'a dying boss', where Steffens told him why corrup-
tion is evil (see Part III, elL VIII, p. 419).
The problems of corruption in post-colonial societies are discussed by S. Andreski,
Tile Afriea PrIllU.",tJII (Micbael Joseph, London, 1968). Although the author discusses
the problem with sympathy and insight, and even ofFers the technical term 'kleptocracy',
his analysis Iacb the depth ofSteft"eus' (ofwhich he seems unaware), and it suffers from
his assumption that in advanced societies corruption is minor in its scale and efFects.
6 It is important to realize that even the pnctic:e of a legislator actively promoting
measures for his direct financial benefit is not necessarily corrupt; this was a common and
accepted state ofaffairs in Victorian England. See R.. A. Lewis, Ed"';" Cud",;,i au the
PU/U H,IIlt" MtIWIIIeIII1832-18S4 (Longman., Green, 1952), ch. 15, 'Reaction" which
 Conclusion: The Future ofScience 419
affairs, or completely cynical, or capable of some sort of double
morality, he may maintain his personal integrity. But more com-
monly, a shadowy awareness that things are not quite as they must
be claimed to be, will force the individual agent to recognize the
possibility of his complicity in something culpable. Such recognition
tends to preserve and intensify the corrupted state of the activity:
fear of exposure comes to dominate the purposes of the agent, and
the group as a whole is held together by mutual blackmail. In such
a situation, the worst elements gain power over the better, and the
performance of the professed social functions is the least of the
considerations affecting individual and collective decisions.
It is easy to see that this sketched analysis applies to cases of
corruption in public life, where the private goals are the venal ones
of personal gain. On the other hand, it is by no means necessary that
every bureaucracy in which the defining functions have been dis-
placed is corrupt; if there is no significant public which had some
trust in it to begin with, there has been no betrayal. But again, it is
possible for officials of a voluntary or political organization to be-
come corrupted without desiring or achieving any personal benefit,
merely by finding it impossible to achieve the purposes of their
members and also finding it impossible to confess their failures.
And most tragically, it is possible for a man to discover in retro-
spect, after years of service to an organization, that he had all
along been corrupt. 7
Cases of corruption in technical projects can be quite straight-
forward: a public contract for a device is sought and procured, on
the basis of promised operating characteristics which the contractor
has neither the ability nor the intention of achieving, but where the
failure of the project will not affect his interests adversely. To pro-
tect his interests, the contractor may find it necessary to corrupt the
State agencies of control, so that they will merely pretend to
or
describes the successful campaign in Parliament to defend the privileges the Lcmdoa
water companies against the needs ofthe population. Even in the early twentieth century
Lincoln Steffens found English politicians who denied the existence of corruption in
England calmly desaibing practices which in America are certainly amsidered corrupt.
In trying to explain the differences between Europe and America he invoked the idea of
the 'old' and the 'new' civilizations, with 'corruption' as a natural historical process. Sec
Autobiography, Part IV, ch. VIII.
7 The type-case of unwitting corruption is Steffens's Captain Schmittberger, a
German immigrant who simply never knew that the policeman's job does fUJI include
protecting rackets and taking bribes, until the Lexow investigation exposed the truth
about the system, to himself. See Autollio"'.'''1, Part II, cbs. XII and XIII, pp. 266-8+
420 Conclusion: The Future ofScience
scrutinize his operations. 8 The situation is similar in practical projects:
the construction of a research empire may be publicly justified by its
function in the solution of urgent practical problems, while the
effective goals of the project are the provision of jobs, the securing of
an academic base, and the production of titles of publications. Even
in academic science, the production and publication of shoddy work
involves an element of corruption, since both author and referee are
participating in a deception, albeit an ambiguous one before a
largely anonymous public. Entrepreneurial science is by its very
nature corrupt in this sense, and an immature field in a hyper-
trophic state can scarcely avoid corruption. We notice, however,
that as we move away from the straightforward situations where
bank-notes are passed in return for favours, the subtlety and
ambiguity of corruption become more pronounced. Indeed, it is
possible for one person to denounce a project as corrupt, and another
indignandy and sincerely to deny it, the disagreement resting on
questions of whether there is an interested public, a trust, and a
betrayal.
A failure to come to terms with the new problems resulting from
the industrialization of science can bring about a very subtle but
none the less corrosive form of corruption within science as a whole.
For science, as a part ofacademic scholarship, has long claimed to be
discharging a public trust in the advancement and diffusion of
knowledge; and it has claimed a variety of privileges and immunities
for its members (not shared by other teachers or research workers)
on the basis of its ethic of autonomy and integrity. These claims are
different in character from those claimed by a learned profession,
and in some ways more extreme: a professional is expected to use
his judgement in solving the problem set by the welfare of the
client, while the scientist or scholar claims the freedom to choose the
problem itself. If it were possible to make a neat separation between
the sections of the scientific or academic community which are
devoted to scientific problems on the one hand, and technical and
practical problems on the other, then each section could develop its
distinct identity, with appropriate public justification of its position
and appropriate methods of social behaviour. But the industrializa-
tion of science brings the different sorts of problems into ever closer
connection, in institutions and in the work of individuals. There is
naturally a great temptation for the leaders of science to attempt to
• See Nieburg. /" tht N"mt ofSdmtt, for a discussion of this process.
 Conclusion: The Future ofScience 42 1
gain the best of both worlds for as long as possible: to extol the
virtues of the free search for truth to one audience, and to promise
useful services to another. Except in those happy coincidences when
the different criteria ofvalue yield identical choices in a field, such a
situation is liable to produce corruption. The most easily identi-
fiable situation with such tendencies is what Americans call 'bootleg'
research, where resources obtained for one project are partly, at
least, diverted to another of more interest to the investigator. If this
is something that 'everyone does' in a particular community, the
equivalent of 'fiddling', then it is not corrupting to those involved. 9
On a large scale, however, as in big entrepreneurial science it can
have serious consequences.
A more subtle but more dangerous sort of corruption can occur
when the balance of real and professed final causes is tipped the
other way: when scientists claim the privileges appropriate to the
heirs of Helmholtz, while accumulating personal wealth and institu-
tional power through the regular contracting of mission-oriented
research. Again, the corrupting effects of such a situation may be
latent, until it is exposed and challenged. The natural response is
then to hide what can still be hidden and to explain away what
cannot. Up to now, such exposures, and their attendant crises, have
occurred only in connection with dirty science in American uni-
versities. Once the dangers of this situation were recognized, the
response of many leaders ofthe scientific community was admirabl~.
They disengaged from the State in military scientific establishment,
doubtless at considerable personal cost, when the Vietnam war
became politically and morally indefensible, and even before militant
students forced the issue at leading universities. However, such a
move, although welcome and heartening in itself, does not resolve
the underlying dilemma of the external relations of industrialized
science, with its tendency to corruption. to
9 See D. s. Greenberg, '''Boodeggingn: it Holds a Firm Place in Conduct ofResearch',
Snmet, 153 (19 September 1966), 848-9. With cbaracteristically American sophistication,
some large industrial research laboratories become worried if their scientists work only
on the projects formally agreed upon.
10 Some scientific communities maintain their independence and integrity by
utonishingly direct means. In Japan, physicists who associate themselves with the Japan
Defence Agency are ostracized by the Japan Physical Society, not being permitted to
Pl\*Ilt papers at its conferences. Military personnel sent to graduate schools by the
armed forces are failed on their exams, either on entrance or on completion of their
coone. In these circumstances, it is not surprising that 'Japanese defence oflicials have
also privately admitted to American colleagues that they have difficulty getting scientists
IS-S.te.S.P.
422 Con&lwion: The Future ofScience
The State, and industry, need an expertise more sophisticated and
prestigious than can be provided by narrowly technical institu-
tions and personnel; it must come from the world of science based
on academic institutions. I I And if it is to continue to receive support
on the present large scale, science must provide this service. To do
so effectively, in the Anglo-American institutional structure, re-
quires science to maintain a plausible semblance of its autonomy and
integrity; yet the autonomy of science cannot be more than a
semblance, especially as its accountability to its paymasters becomes
increasingly close and obvious. The world of science will then need
to half-believe itself to be still academic, free and autonomous, while
half-knowing itself to be industrialized, dependent, and responsible
to the State and industry. The traditional professed functions of
science are internal and under its control: a means to the ultimate
goal of the advancement of knowledge, as conceived and guided by
itself. But the actual goals of the work are increasingly subordinate
to functions that are externally defined: a means to the fostering of
civil and military industry, through the provision of particularly
applicable results and trained experts, on demand. When and
whether this divergence will become a part of the self-consciousness
of science, and by whom it could then be considered as a betrayal of
a public trust, depends entirely on the complex circumstances of
history and society. In this connection, one may revise Lord Acton's
aphorism, 'all power corrupts, and absolute power corrupts abso-
lutely', and substitute, 'responsibility tends to corrupt, and respon-
sibility without power corrupts absolutely'.12
Critical Science: Politics and Philosophy
We can now permit ourselves some final speculations on possible
ttends in the future of the natural sciences. The process of industri-
to pcrfonn defence research'. See P. M. Boffey, 'Japan (I): On the Threshold of Big
Science ?', Scim&t, 16, (1970), 31-5.
II 'Precisely the indiscipline ofrelatively free intel1ectual activity attracts the powerful
u guaranteeing the relative disinterestedness and novelty they hope to find in the ideas
of intel1ectual counse11on. It is one of the ironies of our time that so many intellectuals
strive to identify with the penpectives of kings, whilst their rulers value them for their
activity as philosophers.' See N. Birnbaum, 'On the Idea of a Political Avant-garde in
Contemporary Politics: the Inte11ectuals and Technical Intelligentsia', Praxis (Zagreb,
1969), Nos. 1-2, p. 343·
12 This aspect of corruption is discussed in my papers on 'Power, Responsibility,
Answerability'. The context there was the problem of participation in university
government; but it could be interesting to relate it to the position ofscientists responsible
to the State described by the Amaican c1ich6, 'on tap but Dot on top'.
 Conclusion: The Future ofScience 423
alization is irreversible; and the innocence of academic science
cannot be regained. The resolution of the social problems of science
created by its industrialization will depend very strongly on the
particular circumstances and traditions of each field in each nation.
Where morale and effective leadership can be maintained under the
new conditions, we may see entire fields adjusting successfully to
them, and producing work which is both worthwhile as science and
useful as a conttibution to technology. Recruits to this sort ofscience
will see it as a career only marginally different from any other open
to them; and it is not impossible for men of ability and integrity to
rise to leadership in such an environment. This thoroughly in-
dusttialized science will necessarily become the major part of the
scientific enterprise, sharing resources with a few high-prestige
fields of 'undirected' research, and allowing some crumbs for the
remnants of small-scale individual research. A frank recognition of
this situation will help in the solution ofthe problems ofdecision and
conttol. Since the criteria of assessment of quality will be heavily
biased towards possible technical functions of results, they will
thereby be more easily applied, and less subject to abuse, than those
which are based on the imponderable 'internal' components of
value.
Thus, provided that the crises in recruitment and morale do not
lead to the degeneration and corruption of whole fields, we can
expect emergence of a stable, thoroughly industrialized natural
science, responsible to society at large through its conttibution to
the solution of the technical problems set by industry and the State.
Scientists, and their leaders and institutions, will be 'tame': accept-
ing their dependence and their responsibilities, they will be unlikely
to engage in, or encourage, public criticisms of the policies of
those institutions that support their research and employ their
graduates. Such a policy of prudence is not necessarily corruption;
whether it becomes so will depend on many subde factors in the
self-consciousness of this new sort of science, and the claims made
to its audiences. But not all the members of any group are easily
tamed, and the emergence of a 'critical science', as a self-conscious
and coherent force, is one of the most significant and hopeful
developments of the present period.
There have always been natural scientists concerned with the
sufferings of humanity; but with very few exceptions they have
faced the alternatives of doing irrelevant academic research to gain
424 Ctm&lusion: The Future ofScience
the leisure and freedom for their social campaigns, or doing applied
research which could benefit humanity only if it first produced
profits for their industrial employer. The results of pharmaceutical
research must pass through the cash nexus of that industry before
being applied, and that process may be an unsavoury one. Only in
the fields related to 'social medicine' could genuine scientific research
make a direct conttibution to the solution of practical problems, of
protecting the health and welfare ofan otherwise defenceless public.
Now, however, the threats to human welfare and survival made by
the runaway technology of the present provide opportunities for
such beneficial research in a wide range of fields; and the problems
there are as difficult and challenging as any in academic science.
These new problems do more than provide opportunities for scien-
tific research with humanitarian functions. For the response to this
peril is rapidly creating a new sort of science: critical science. In-
stead of isolated individuals sacrificing their leisure and interrupting
their regular research for engagement in practical problems, we
now see the emergence of scientific schools of a new sort. In them,
collaborative research of the highest quality is done, as part of
practical projects involving the discovery, analysis and criticism of
the different sorts of damage inflicted on man and nature by run-
away technology, followed by their public exposure and campaigns
for their abolition. The honour of creating the first school of 'critical
science' belongs to Professor Barry Commoner and his colleagues at
Washington University, St. Louis, together with the Committee for
Environmental Information which publishes Environment. 13 For
some years a Society for Social Responsibility in Science, based in
America but with members and branches overseas, was the main
voice of conscience in science; recently a British Society for Social
Responsibility in Science, with a rather broader base among the
leaders of the national scientific community, has been fonned. As
such societies gain sttength and influence, the success of the St.
Louis school of critical science should soon be emulated elsewhere.
The problem-situations which critical science investigates are not
the result of deliberate attempts to poison the environment. But
they result from practices whose correction will involve inconveni-
ence and money cost; and the interests involved may be those of
powerful groups of firms, or agencies of the State itself. The work
I J The first statement of 'aitical science' as distinct from ecological concern is
B. Commoner, Stimtt ,,1Ul Swvi.l (Gollanc:z, London, 1966).
 Conclusion: The Future ofScience 42S
of enquiry is largely futile unless it is followed up by exposure and
campaigning; and hence critical science is inevitably and essentially
political. 14 Its style of politics is not that of the modem mass move-
ments or even that of 'pressure groups' representing a particular
constituency with a distinct set of interests; it is more like the
politics of the Enlightenment, where a small minority uses reason,
argument, and a mixture ofpolitical tactics to arouse a public concern
on matters of human welfare. IS The opponents of critical science
will usually be bureaucratic institutions which try to remain faceless,
pushing their tame experts, and hired advocates and image-
projectors, into the line of battle; although occasionally a very
distinguished man is exposed as more irresponsible than he would
care to admit. 16
In the struggles for the exposure and correction of practices
damaging the environment the role of the State is ambiguous. On
the one hand, every modem government is committed in principle
to the protection of the health of its people and the conservation of
its natural resources. But many ofthe agencies committing the worst
outrages are State institutions, especially the military; and in any
event the powerful interests which derive profit or convenience from
polluting and degrading the environment have more political and
economic power than a scattering of 'conservationists'. It some-
times occurs that two State agencies will be on opposite sides of an
environmental sttuggle; but the natural tendency ofregulatory agen-
cies to come under the control of those they are supposed to regulate
can make such a struggle a one-sided affair.
The presence of an effective critical science is naturally an
embarrassment to the leadership of the responsible, industrialized,
14 The most comprehensive analysis of 'aitical science' yet published is MuNichoJson.
Tile EmJirtnmle1ltlll Revolution (Hodder & Stoughton, London, 1970). He is mainly
concerned with 'conservation', but his healthy approach to modem bureauaatic politics
is developed in his earlier book, Tile System (Hodder & Stoughton, London, 1967;
McGraw-Hill. 1969).
15 Support for this new style of politics bu come from the Duke of Edinburgh.
Speaking on the B.B.e. programme '34 Hours', on 17 February 1970, he discuaed the
proposition that 'tough action against the poisoners and wrecken' was essential for the
promotion ofconservation and the abatement of pollution, and said that people must 'be
ruder and more direct to the people in political authority'.
16 On 22 March 1966, for example, the President of General Motors appeared before
a Senate hearing to apologize to Mr. Ralph Nader, following revelations that General
Motors' lawyers had hired an investigator to unearth details about Nader's private life.
Nader's analysis of the defects of the 'Conair' car was aJSting General Moton a lot of
money.
 426 Con&lusion: The Future ofS,imce
tame scientific establishment. l ? Their natural (and sincere) reaction
is to accuse the critics of being negative and irresponsible; and their
defensive slogan is along the lines of 'technology creates problems,
which technology can solve'. This is not strictly ttue in all cases,
since nothing will solve the problems of the children already killed
or deformed by radioactive fallout or by the drug Thalidomide. ls
Moreover, this claim canies the implication that 'technology' is an
autonomous and self-correcting process. This is patent nonsense.
We have already seen that a new device is produced and diffused
only if it performs certain functions whereby human purposes can
be served; and if the intended beneficiaries do not appreciate its use,
or if those injured by its working can stop it, the device will be still-
born. The distortions of technological development arise when the
only effective 'purposes' in the situation are those of the people who
believe themselves to derive pure benefit from the innovation. On
the self-correcting tendency of technology, one might argue that no
large and responsible institution would continue harmful practices
once they had been recognized; but this generalization is analogous
to the traditional denial of the cruelty of slavery, along the lines that
no sensible man would maltreat such valuable pieces of property.
And the history of the struggles for public health and against pol-
lution, from their inception to the present, shows that the guilty
institutions and groups of people will usually fight by every means
available to prevent their immediate interests being sacrificed to
some impalpable public benefit. 19 If the campaigns waged by
mtical science come to touch on some issue centtal to the conven-
ience of the State or other very powerful institutions, we may
experience a polarization of the community of natural science,
along the same lines as has already occurred on the Vietnam issue in
some of the human sciences in America. In such a situation, it will
17 Thus N"""e ridiculed the UNESCO amference on the biosphere of September
1968t comparing it to an earlier amference on 'communications satellites and under-
deft10ped countriest. See 'Bandwapn for UNESCO't N.,,,,e, 219 (7 September 1968)t
p. 999. There was DO report on the amference itself:
II Up to the time ofwritingt N"""e maintains a magnificent complacenCYt such u one
bardIy hopes to see in England in i1B epoch of imperial decline. Thus, aiticizing Dr.
Fruer Darling'. fint Reith Lecture. an editorial ~ 'And in spite of quite proper
concern for the need to make ouly decent use of new developments, it is hard to find
contemporary i llustntions of where teclmology has gone astray.' 'No Peace for the
Wicked t t N"""e, 234 (1969), 63 1 •
19 The 'IIDitary movemeni in nineteeDth-century England was involved in such
IttUp through i1B career. For a ample, see It A. Lewis, U",;n Cluulwi tIIUl ,lie
p"jlie HMlII MIIfJaIeIII 1832-S4 (Longmans, Green, 1952).
 Conclusion: The Future ofScience 427
not be possible for a leader of science to be both honest and tame;
and if the establishment of science chooses to serve its paymasters
rather than ttuth, it will be recognizably corrupt.
Such extreme situations may be a long time in developing, if for
nothing else than that critical science is still in its infancy. As it
develops, it will be at risk of encountering many pitfalls, partly
those characteristic of immature sciences applied to practical prob-
lems, and partIy those of radical and reforming political move-
ments. Perhaps the most obvious will be an accretion of cranks and
congenital rebels, whose reforming zeal is not matched by their
scientific skill. But there are others, arising from the contradictory
relations between critical science and the relevant established in-
stitutions of society. As ttue intellectuals rather than a technical
intelligentsia, individual members may find some 'sinecures within
the interstices of bureaucratized intellectual systems'; 20 but there
will need to be some institutions providing a home for the nucleus
of each school, and external sources of funds for research. Hence,
especially as critical science grows in size and influence and society
becomes more sophisticated about the problems of runaway tech-
nology, some accommodation between the critics and the criticized
will inevitably develop. We can even expect to see critical research
being supported, critical slogans being echoed, and leaders of
critical science being rewarded, by institutions whose basic des-
tructive policies still are unchanged. 21 Such phenomena have
already occurred in America, in the politics of race; and on this
issue, where the interests concerned are mainly major institutions
which can hire talented and enlightened experts at will, it is even
more likely. The movement of critical science would then face the
pitfalls of corruption as soon as, or even before, it had skirted those
of impotence. 22 But this is only a natural process, characteristic of
20 ' ••• the intellectuals' distance from certain kinds of material activity, their occupa-
tional repugnance for certain forms of boUl'pOis organizations, their attachment to
abstract versions of bourgeois tradition nther than to the lUb-stratum of bourgeois
activity, their familiar quest for sinecures within the intentices of bureaucratized
intellectual systems, combine to endow them with what was once an anticapitalist and is
now an anti-bureaucratic ethos.' See N. Birnbaum, Ope cit., p. 24+
31 First prize in the 'enlightenment' stakes bu been won by the Mcmsanto C1emical
Company. The Scientific Division of the Committee for Environmental Information
(which publishes EmJirtnmle1lt) included among its members for 1969 Mr. F. D. Wharton,
Jr., St. Louis Development Manager, Life Sciences in the New Enterprise Division,
Monsanto Company.
23 On the corruption of good causes, the classic is G. B. Shaw, M"jor BM••• The
climu of the play comes when Mrs. Baines, the Salvation Army (Ammissioocr who
428 Conclusion: The Future ofScience
all radical movements. It is easy to maintain one's integrity when
one's words and actions are ineffective; but a long period of this can
produce a sectarian or a crank. If one begins to achieve power, and
one's policies affect the interests of many others, one must decide
where one's responsibility lies. If it is to the ideal alone, then one is
set on a course towards tyranny, until overthrown by the host of
enemies one has raised Up.23 And if one accepts responsibility for
the maintenance of a general welfare, including that of one's
opponents, one is on the path to corruption and impotence. This
may seem a gloomy prognosis; but a society which does not present
such hazards to radical movements of every sort is not likely to
retain its stability.
We can expect, then, that the future political history of critical
science will be as complex and perhaps as tortured as that of any
successful radical and reforming movement. But if it does
survive the pitfalls of maturation, and so contributes to the
survival of our species, it can also make a very important con-
ttibution to the development of science itself. For if the style
of critical science, imposed by the very nature of its problems,
becomes incorporated into a coherent philosophy of science, it
will provide the basis for a transformation ofscientific inquiry as
nIDI the shelter in West Ham, tbanb God for the douation of {.s,ooo by Mr. Bodger,
the distiller whose whiskey is the curse of the poor in their care. The Cockney Bill
Walker, whose donation of a guinea in repentance for striking two women bad just
previously been indignandy rejected, utters the significant 'Wot prawce selvytion nab?'
(penguin Boob, 1960; p. 106.) Sec also Shaw's Preface to the play.
:&J A cautionary tale that should be read by aU who are embarking on political activism
baed on 'aitical science' is the play by Ibsen, Tile £""'" oftile PtOple. Superficially, it is
about an honest doctor who is hated by the corrupt forces of his town for his determina-
tion to expose the scandal of polluted waters being used in the town's profitable baths.
But on closer reading, it can be seen that Dr. Stoebnann'. misfortunes were also due to
his own D&i~ and egoism. It is significant that in his own version of the play (Viking
Praa, New York, 1951), Arthur Miller strengthened the 'progressive' message by trans-
ferring the paaap where the public meeting declares Dr. Stockmann to be 'an enemy of
the people'. In his version it comes at the very beginnins of the meeting, before he has
spoken; while in the original it comes after the Doctor'. harangue, amcluding with, 'Let
the whole country perish, let all these people be exterminated.' Mer studying the play
with a class at Harvard, where this modification was discovered, it struck me that a
worthwhile sequel could be written, entided 'The People'. Friend', in which the
entrenehed forces, ifouly a bit lea stupid and venal than in the original, could corrupt the
good Doctor without difticulty. Although Spa resorts are DO longer an important focus of
pollution, it is possible for an 'Enemy of the People' situation to be repeated in any of the
seaside toWDS which dump their nw sewage into proDmity to bathers. See J. A. Wakefield,
'Qean or Dirty Beaches-Which do you Prefer?' Y"" ErrvirtnIIMIII, No. 1 ~mtel'
1969), pp. 29-3 1•
 Conclusion: The Future ofScience 429
deep as that which occurred in early modem Europe. The
problems, the methods, and the objects of inquiry of a matured
and coherent critical science will be very different from those of
academic science or technology as they have developed up to now;
and together they can provide a practical foundation for a new
conception of humanity in its relations with itself and the rest
of nature.
The work of inquiry in critical science involves an awareness of
craft skills at all levels, and tile conscious effort of mastering new
skills. The data itself is obtained in a great variety of ways, from the
laboratory, from the field, and from searching through a varied
literature, not all of it in the public domain. Much of it lacks sound-
ness, and all of it requires sophisticated and imaginative treatment
before it can function as information. Indeed, since the problem-
situations are presented in the environment, and much of the
crucial data must be produced under conttolled conditions in the
laboratory, work in critical science may overcome the dichotomy
between field-work and lab-work which has developed in science,
even in the biological fields, over the past century. In the later phases
of investigations of problems, the same challenges of variety and
novelty will always be present. The establishment of the sttength
and fit of each particular piece of evidence is a problem in itself;
and the objects of inquiry (including the measures of various effects
and processes, as well as conventional standards of acceptability in
practice) are so patently artificial, that one is in little danger of
being encased in them as a world of common sense. The establish-
ment of criteria of adequacy for solved problems is possible, for the
work will frequently be an extension and combination of established
fields for new problems, and so critical science can escape the worst
perils of immaturity. Also, any critical publication is bound to be
scrutinized severely by experts on the other side, so high standards of
adequacy are required because of the political context of the work.
Indeed, a completely solved problem in critical science is more
demanding than in either pure science or technology. In the former,
it is usually sufficient to obtain a conclusion about those properties
of the artificial objects of inquiry which can be derived from data
obtained in the controlled conditions of experiment; in the latter it
is sufficient for an artificial device to perform its functions without
undue disturbance by its natural environment; while here the com-
plex webs of causation between and within the artificial and natural
 430 Ctm&lusitm; The Future ofScience
systems must be understood sufficiently so that their harmony can
be maintained.
The social aspects of inquiry in critical science are also conducive
to the maintenance of its health and vitality, at least until such times
as the response to its challenge becomes over-sophisticated. The
ultimate purpose which governs the work is the protection of the
welfare of humanity as a part of nature; and this is neither remote,
nor vulgar. Critical science cannot be a permanent home for career-
ists and enttepreneurs of the ordinary sort; although it may well use
the services of bright young men intending eventually to serve as
enlightened experts. Those who want safe, routine work for the
achievement of eminence by accumulation will not find its atmos-
phere congenial; for its inquiries are set by a succession of problem-
situations, each presenting new challenges and difficulties. Hence
although critical science will doubdess experience its periods of
turbulence, political and scientific, it is well protected from stagna-
tion and from the sort ofcreeping corruption that can easily come to
aftliet industtialized science.
Finally, the objects of inquiry of critical science will inevitably
become different from those oftraditional pure science or technology,
for here the relation of the scientist to the external world is so
fundamentally different. In traditional pure mathematical-experi-
mental natural science, the external world is a passive object to be
analysed, and only the more simple and abstract properties of the
things and events are capable of study. In technology, the reactions
of the unconttolled real world on a constructed device must be
taken into account, but only as perturbations of an ideal system; the
task is to manipulate it or to shield the device from its effects. But
when the problem is to achieve a hannonious interaction between
man and nature, the real world must be tteated with respect: both
as a complex and subtle system in its own right, and as a heritage of
which we are temporary stewards for future generations. Hence,
even though studies of our interaction with the environment will
necessarily use all the intellectually constructed apparatus of dis-
ciplined inquiry, their status and their content will inevitably be
modified. They will be more easily recognized as imperfect tools,
with which we attempt to live in harmony with the real world around
us; and although this attitude may seem to conduce to scepticism,
it will be the healthy one which recognizes that genuine knowledge
arises from lengthy social experience, and that such knowledge
 Conclusion: The Future ofScience 43 1
depends for its existence on the continued survival ofour civilization.
The objects of inquiry themselves will include final causes among
their essential attributes, not merely the limited functions appro-
priate to technology, but also the judgements of fitness and success
already developed in classical biology and ecology. All this is work
for the future; but if it is successful, the opposition between
scientific knowledge and human concerns, characteristic of the
sciences derived from the dehumanized natural philosophy of the
seventeenth century, will be overcome.
With a new conception of the practice of science, there will come
a new conception of the history of science and of the meaning of the
scientific endeavour. It is possible, and it has been natural, to
reconstruct the history of scientific inquiry as a success-story
leading up to the ttiumph of the matured academic mathematical-
experimental natural science of the later nineteenth century. One
can identify the historical moments, and the great men associated
with them, when the very conception of the inqUiry itself was
significantly advanced. Thus, the origins of our sort of science are
rightly located. in the earlier Greek civilization, when attempts were
made to account for the world of sense experience without invoking
personified divine agents. The heritage ofthe so-called 'pre-Socratic'
philosophers was further developed by Aristotle, who not merely
conducted disciplined inquiries over almost all areas of human ex-
perience, but also showed that such disciplined inquiry has conceptual
and methodological problems that can and should be investigated. In
a parallel tradition, the idea of mathematics as a body of proved
results about conceptual objects developed to full maturity in the
achievements of Euclid and Archimedes. The next great advance
(according to this interpretation) came many centuries later, 'when
the pioneers of the 'mechanical philosophy' of the seventeenth
century achieved a powerful synthesis of experience and reason.
GaliIeo's slogan, 'sense experience and necessary demonstration',
stood for his appreciation of the need for closely controlled experi-
ence, which could serve as evidence of the appropriate strength, in
an argument cast in mathematical language. All that was required to
complete the body of methods of classic academic science was the
development of institutions for organized, co-operative research;
and this came by natural evolution through the nineteenth century.
The dominant world-view of matured academic science was atom-
istic in several important senses. The real world underlying our
432 Ctm&lusion: The Future ofScience
sense experience was assumed to be devoid of the characteristically
human attributes of intellectual and spiritual reality, and of value;
final causes were excluded from scientific explanation; and all
efficient causes were to be reduced to the material cause of insensibly
small brute matter in motion. Correspondingly, knowledge itself
was atomic: the achievement of individual facts about the external
world, isolated from any philosophical and social context, was
considered possible and valuable. This approach to natural know-
ledge achieved magnificent successes in many fields, and was also
appropriate for the development of successful large-scale research.
It was natural to suppose that this particular style of scientific
enquiry could be successfully extended to all disciplines, and that it
was internally stable. But both these optimistic assumptions proved
incorrect. Ineffective and immature fields can be hindered rather
than helped by a mechanical imitation of those whose objects and
appropriate methods are very different; and the pretence of this sort
of maturity only adds to the hazards of applying such fields to
practical problems. On the other hand, academic natural science has
been transformed by its very successes into industrialized science;
and the unexpected problems and abuses of this new sort of science,
ranging from shoddy science to runaway technology, present
threats to the survival of science and of our whole civilization.
With the new perspective gained from our recent experience, we
can look again at the long history of the human endeavour of under-
standing and controlling the external world. We can now see a
positive significance in events and tendencies that have hitherto been
considered as unfortunate aberrations. The dominant traditions in
academic science have developed out of conflict with other styles of
scientific work; and it can be distinguished from them by its objects
of inquiry, its methods of work, and its social context. For 'our'
science, the real world is devoid of human and spiritual qualities;
and the scientist studies the smaller aggregations of matter, con-
sidering the most simple and mathematical properties that suffice
for the successful investigation of problems. Its approach to its
materials is appropriately depersonalized; any 'deeper' meaning
that might be thought to inhere in its results is rigidly segregated
from his reporting, and is left to amateur speculations. As a social
activity, this science is necessarily elitist, presupposing a lengthy
course of training and indocttination for which only a minority have
an appropriate cultural background.
 Conclusion: The Future 0/ Science 433
To the extent that the ttaditional history ofscience has considered
these aspects of scientific inquiry, it has been embarrassed by the
presence of traditions and tendencies that achieved success in 'our'
terms in spite of radical differences in one or more respects from the
recently dominant academic style. The roots of astronomy in
astrology, and of chemistry in alchemy, are cases in point. Some of
the immortal ancestor-figures of the modem discipline are revealed,
on unbiassed inspection, to have seen their work as conttibuting to
what is now regarded as pseudo-science: Ptolemy and Tycho for
astronomy, and Paracelsus and Glauber for chemistry. Indeed,
when we look more closely at the period of the later sixteenth
century, when the arts and sciences were developing quite rapidly
he/ore the incursion of the 'new philosophy' of dead matter, we find
the very greatest scientists participating in the world-view of an
animated nature: Gilbert investigating magnetism in the attempt to
prove that the earth is the embodiment of the anima mundi, Kepler
searching (with all rigour) for the divine harmonies of the celestial
realm, and Harvey using 'spirit' and the macrocosm-microcosm
analogy to guide his anatomical and physiological researches. Z4
It would be very misleading to imagine a simple succession oftwo
sorts of science, each unified and coherent in itself, first that of the
'animated' world and then (since the seventeenth century) that of
the 'dehumanized and disenchanted' world. History is more complex,
and more interesting, than that; and within the 'old' conception of
science there were many different tendencies in the interpretation
of its appropriate objects, methods and social functions. I have
previously referred to a 'romantic' philosophy of nature providing
the vehicle for a politically radical folk-science that challenges the
academic science of its time. In this tradition, the study of nature is
explicitly seen as a social and also spiritual aet; one dialogues rather
than analyses; and there is no protective cover of belief in the
'neutrality' or 'objectivity' ofthe style adopted. Such a philosophy of
nature will become articulated and advanced, as part of a general
radical reaction against a formal, dry or bureaucratic style pervading
social or cultural life. Looking back into history, we can find a
similarity of doctrine or style, and sometimes a linking ttadition, as
24 Gilbert makes his programme plain; see tie M"pete Book V, ch. 12: 'The magnetic
force is animate, or imitates a soul; in many respects it surpuses the human soul while
that is united to an organic body! (b'. P. F. MotteIay; Dover, New York, 1958, p. 308.)
Kepler is well-known; and for Harvey, see W. Pagel, WillUJ", H1lrW1'1 BioloF'" Itle.,
(Buel and New York, 196'7).
 434 Ctm&lusion: The Future ofScience
far back as the Taoists of ancient China, through St. Francis of
Assisi, to Paracelsus, William Blake, and Herbert Marcuse. 25
Not every one of these figures would claim to be a natural
scientist of any description; but as philosophers, poets or prophets,
they must be recognized as participating in and shaping a ttadition
of a certain perception of nature and its relation to man. Granted all
the variety of their messages and styles, certain themes recur. One
is the 'romantic' striving for immediacy, of contact with the living
things themselves rather than with book-learned descriptions.
Another is 'philanthropy'; the quest is not for a private realization,
but for the benefit of all men and nature. And, related to these is a
radical criticism ofexisting institutions, their rules and their person-
nel. Looked at from the outside, each upward thrust of the romantic
philosophy of nature is doomed to failure. Mankind. will not be
transfigured overnight; and the romantic style has its own destruc-
tive conttadietions. Whereas the 'classic' style degenerates gradually
into an ossified form and a sterile content, the 'romantic' style goes
off much more quickly, through chaos of form and comfption of
content. But this study of ours has shown that even in disciplined
scientific inquiry, the categories of 'success' and 'failure' are neither
so absolutely opposed, nor so assuredly assignable in particular cases,
as the ttaditional ideology of science assumed. And the failure to
achieve Utopian dreams, in science as well as in social reform, is
not at all the same thing as futility.
The dreams of the romantic, philanthropic, radical philosopher-
prophets cannot move towards realization by the accumulation of
facts or of battalions. Rather, they exist through a discontinuous,
perhaps erratic, series of crises and responses. Sometimes they have
the good fortune of producing a creative tension in a man brave
enough to attempt the synthesis of a prophet's vision with a world
as On Taoism, see J. Needham, Seienee .rul CiviliSlltitm in Chi"", 2, 88-132. In his
magisterial fashion, Needham provides more materials on the analogous movements in
early modern Europe than is available in any general history ofscience. For a discussion
of the limitations of his view, see note 35 on p. 394 above. See Lynn White, Jr., 'The
Historical Roots ofourEcological Crisis', MII&hine ez/)eo (M.I.T. Press, 1968),Chapter 5,
for St. Fnncis.
Fnncis A. Yates, in 'The Hermetic Tndition in Renaissance Science' in ~t, StilfU'e
all HilttW:! ill tile ReuiSSllll&e, ed. C. S. Singleton Oohn Hopkins Press, Baltimore, 1968),
discusses the 'Rosicrucian' style ofscience in considerable depth. The theme.of 'philan_
thropy' is most clearly developed in the German alchemical philosophers in the Paracel-
Iian tradition; and their influence on Francis Bacon is clear.
On Marcuse, see his One-Dimmsitnull M." (Bcac:on Press, Boston, 1964).
 Conclusion: The Future ofS&imte 435
managed by priests. He too will fail, almost certainly; some problems
are insoluble. But his message, perhaps in a particular science or
walk of life, perhaps of a generalized wisdom, will speak to men in
later ages, coming alive whenever it has insights to offer. In this
present period, we may find Francis Bacon speaking to us more than
Descartes the metaphysician-geometer or Galileo the engineer-
cosmologist. As deeply as any of his pietistic, alchemical forerun-
ners, he felt the love of God's creation, the pity for the sufferings of
man, and the striving for innocence, humility, and charity; and he
recognized vanity as the deadliest of sins. 26 To this last he ascribed
the evil state of the arts and sciences:
For we copy the sin of our first parents while we suffer for it. They
wished to be like God, but their posterity wish to be even greater. For we
create worlds, we direct and domineer over nature, we will have it that all
things are as in our folly we think they should be, not as seems fittest to the
Divine wisdom, or as they are found to be in fact. 27
The punishment for all this, as Bacon saw it, was ignorance and
impotence. It might seem that the problem is different now, for we
have so much scientific knowledge and merely face the task of
applying it for good rather than evil. But Bacon assumed his readers
to believe themselves in possession of great knowledge; and much of
his writing was devoted to disabusing them of this illusion. Perhaps
the daily reports of 'insufficient knowledge' of the effects of this or
that aspect of the rape of the earth, and our sense of insufficient
understanding of what our social and spiritual crises are all about,
indicate that in spite of the magnificent edifice of genuine scientific
knowledge bequeathed to us, we are only at the beginning oflearning
the things, and the ways, necessary for the human life.
Bacon was a shrewd man, fully sensitive to the weaknesses of the
human intellect and spirit. He was aware of the superficiality of
ordinary thought and discourse, at whatever educational level; and
he also distrusted the extraordinary enthusiast, in religion or politics,
for the damage he could cause. His life's endeavour was to overcome
26 For a detailed interpretation ofBacon'. programme for scipnce in terms of. vision
of moral and spiritual reform, see J. R. Ravetz, 'Francis Bacon and the Reform of
Philosophy', in Stimee, Mediane MIll SDtin;yi" tileReuiSllMee (Walter Pagel FestICbrift),
cd. A. Debus (University.ofChicago Press and Oldboume Press, London, 1972). This is
an elaboration of certain themes in Benjamin FarriDgton's, Tile PJaloSlllIl;y of p,,,.;,
Bilton, and I am indebted to him for my tint iDsights into this aspect of Bacon.
27 See Tile NlltWtU "rul Ezpm1llllll1ll HUttW,ftW tile F~ qfPIIiIoIfJ,II, (WtWa.
vol v; traD8Iation p. I~).
 436 Conclusion: The Future ofScience
this conttadietion somehow, and to bring about a ttue and effective
reformation in the arts and sciences of nature. For him, this was a
holy work, a work of practical charity inseparable from spiritual
redemption. 28 His audience was inevitably among the literate; and so
he ttied, by scattering hints and half-concealed invitations, to call
together his brothers, who would gendy and silendy show by their
example that a good and pure way into Nature is also the practically
effective way. Of course he failed, in his philosophical reform as in
his political career. There was no English audience for his particular
message during his lifetime, and at his death he was alone and
neglected.
Shortly after his death, however, there was a stining; and Bacon's
message of 'phi1anthropic' science began a career of its own. For a
while, his followers knew what he was about; but with the passage
of decades and disillusion, this was forgotten, and only the vulgar
fact-finding Bacon survived. Yet when we now come back to read
Bacon, perplexed and worned as we are by the sudden ttansfor-
mation that science has wrought upon itself as well as upon the
world, we can find relevance in passages like the following:
Lastly, I would address one general admonition to all; that they con-
sider what are the true ends of knowledge, and that they seek it not either
for pleasure of mind, or for contention, or for superiority to others, or for
profit, or fame, or power, or any of these inferior things; but for the
benefit and use of life; and that they perfect and govern it in charity. For
it was from lust of power that the angels fell, from lust of knowledge that
men fell; but of charity there can be no excess, neither did angel or man
ever come in danger by it. 30
al See the MellillllUmes SIIerM (WtWis, voL vii; translation pp. 243-4). Bacon contrasts
the miracIa ofpunishment wrought by the prophets ofthe Old Testament, with those of
Jesus: 'Jesus was the Lamb of God, without wrath or judgment. All his miracIa were for
the benefit of the human body, his doctrine for the benefit of the human soul' After a
list of instances, Bacon comments, 'There was no miracle of judgment, but all ofmercy,
and all upon the human body.' Later, in the essay 'Of Hypocrites', he comments, 'The
way to convict • hypocrite therefore is to send him from the worb of sacrifice, to the
works of mercy. Whence the text: P"e relititm MUlllllllefiJetl1leftWe Coil "rullM F"/Mr
;s I.S, 10 tUIIM tWpluuu tllUllI1itlo",s ;n 11Ieir "J!Iie'itm •••' (p. 249).
at On the influence of Bacon, see Qwles Webster, S.",.l H.,liIJ "rul'M AtlWlnte-
,.", ofu.";,,, (Cambridge University Press, 1970).
JO Bacon, Till GrMlI",,1IW1JIitnJ, Pre/M' (Wora, vol iv; tnnsJation pp. 20-21).
 INDEX OF NAMES

Actoo, Lortl, 1330, 422 Boffey, P. M., 490, 422n

Adler, E. P., 336n Bok, B. J., 60n
Apssi, J., 172ft Boltzmann, L., 105n
Allardt, E., g6n Boring, E. G., 267ft, 3730, 379Jl
Allen, J., 6411 Born, H., 310n
Alperowitz, G., 64ft Born, M., 31011
Andreski, S., 418n Boa, H. J. M., 299D
Arago, D. F. J., 227 Bottle, R. T., 470
Archimedes, 43 1 Boughey, K., 99D
Arendt, H., 363n Bowden, B. V., 33011
Aristotle, 17, 116-8, 127ft, 1580, 240, 298, Boyer, C. B., 2480, 249
319,43 1 Boyle, Re, 98, 188, 2O.fD
Ashby, Sir E., 29 Bndbury, S., 197ft
Avogadro, A., 2O.fD Bny, J., 340D
Ayres, R. V., 36g0, 399n Bridgman, P. W., 147ft, 202Il
Britten, J., 32411
Babbage, c., 310ft Brock, W. H. 310ft
Bachelard, G., 19011 Brodrick, J., 64ft
Bacon, F., 21, 63, 94t 95, log, 1110, 149n, Bronowski, J., 6sn
16g0, 1700, 308, 375, 376, 434, 435, Brookes, K. J. A., 216n
436 Brouwer, S., 154Jl
Bacoo, R., 399ft Brown, E., 66ft
Bailey, B., 353n Buck, P., 64n
Bain, A., 43n Buckle, H. T., l68n
Barbcrini, M., 391n Bunge, M., 74Dt 31 7
Barzuo, J., 56 Burckhardt, J., 26n
Bell, E. T., 154D, 210n Burrow, J. W., 38gB
Ben-David, J., 430, 67n, 226n, 2860, 309ft
Bendix, R., 350ft Cajon, F., 220
BenD, A. Wedgwood, 363 Calder, N., 363n
Berkeley, G., 2190, 220, 221, 222, 263 Campbell, N. R., xxvi, 84n, 214n
Bemal, J. D., 312 Cardano, G., 248n
Bernoulli, J., 248 Cardwell, D. S. L, xxvii, 44D, 75D, 135n,
Berthollet, M., 226 ~3Jl, ~5D, 206n, 2SOD, 32m
Berzelius, J. J., 2O.fD CarIsoo, E. A., 206ft
Besicovitch, A. S., xxvi, xxvii, 94ft Camot, S., 228, 270
Beveridge, W. I. B., 7sn Carr, E. H., 129ft
Biot, J. B., 1250, 1350, 227, 256 Carter, L. J., 114ft
Birge, R. T., 122n Cauchy, B.tm A. L., 222, 223D, 256
Birnbaum, N., 422 0, 42711 Cavendish, H., 250
Black, J., d19, 2030, 250 Caws, P., 171n
Blackett, P. M. S., 6411 CJwlwick, E., 4180, 426D
Blake, W., 434- Cbemiak, S., 114ft
Blau, P. M., 341n CWde, V. G., 140ft
43 8 Index ofNames
Oampoli, 3910 Einbinder, K, 207D
CJagett, M., 94ft, 3930 EiDstein, A., 65, 660, 1350, 186D, 263n
Clark, Sir G., 34711 264,388
Clark, J. T., 21911 Eliade, M., 144ft
Cifford, W. K., log ElbDa, Y., 94ft
Coddington, A, DVii. 84D, 123n Ellis, B., 2100
Cohen, H., 3410 Ellis, N. D., xxvii, 4311, 33111, 332n, 417n
Cohen, I. B., g8n Ellis, R. L., 630
Coleman,. J. S., 7911 Elton, G. R., 99
CoIliDs, J., 249 Emmett, D., 29 10
Commoner, B., xxviii, 424 EDt, P., 27411
Compton, K., 129ft Euclid, 431
Comte, A., 3870 Euler, L., 219
Conant, J. B., 500, 1610 Ewald, P. P., 196n
Conclorcet, M.",u M, 3870
Cook, F. J., 36311 Fahie, J. J., 64n
Copernicus, N., 194ft, 262, 411 Farber, E., IIID
Countryman, R., 95Jl, 186n, 1960 Farrington, B., 1490, 308D, 435n
Cnnefield, P., 47ft Farris, R. E. L., 3680
Cresswell, H. B., 1000 Feig), H., 213D
Crombie, A c., 1110 Ferrari, 248D
Crone, B., 3590 Finlaisson, J., 4J1l
Crosland, M. P., 1160, u6n, 25 10 Fishbein, M., 48D
Crowther, J. G., 416n Fisher, C. S., 229ft
Fisher, R. A., 79ft, 92D
DUnton, F. S., 2CJ9Il, 400D Fleming, D., 28J1l
Dalton, J., 227, 370 Fourier, J. B. J., 135Jl, IgoD, 213D, 227, 270
Darling, F., 426D Fox, R., 67n, 226n, 263D
Darwin, c., 37911 Fnncis, H. W. S., 3510
Davy, 8., 13SDt 2M Fnncis of Assisi, S.m" 434
Dabus, A G., 3920, 43SD Fnnkel, E., 2560
de Grazia, A., 38gD Frankfort, K, 238D
de Morgan, A., 26m Freedman, P., 75D
Derry, T. IC., 360n Fresenius, 30911
Dacartes, R., 26, 63, 16gB, 1700, 194ft, Fresnel, A J., 227, 270
305, 213ft, 21 7, 320, 345, 39 1n, 435 Freud, S., 379D
Deutsch, M., IUD Freudenthal, H., 256n
Devons, E., 381 n, 3g6n Fuller, R. W., 33SD
Dobrzycki, J., 202Il Fullmer, J. Z., 26m
Dopn, M., g6n
Dolby, R. G. A., xxvii, 16811, 210n, 25On, Galileo, 18, 19, 63, 64n, 73, 98n, IIID,
~521l 123, 124, 127, 137Dt 138, 194ft, 205,
Donnay, J. D. H., 196n 206n, 218n, 219, 249, 262, 266n, 320,
Doria, P., xxvi 365, 39 1n, 413, 414, 43 1, 435
Doty, P., 28711 Galois, E., 228, 270, 411
Drake, S., 98n, 3910 Galton, Sir F., 92n, 37911
DufF, Sir J. F., xxvi Galvani, L., 39, 41 4
Duhem, P., 1270 Garvan, A. N. B., 197D
Dulcmg, P. L., 2~7 Gasendi, P., 194Jl, 219
Duma, J. B. A, 26gn, 27 1ft Gay-Lussac, J. L., 227
Gerusi, J., 294ft
EcIcIIteiD, L, 137ft Gerhardt, C. F., 270
Edinburgh, Ollie of, 42SD Gilbert, L. W., 251n
Eichmann, A., 4 1 5 Gilbert, W., 4 11 , 433
 ltulez ofNfltlles 439
Gillispie, C. c., 2OD, 392D, 3930 Hooybu, Re, IIID
Gilson, E., 63D Houghton-EftIIIt W., 3358
Glass, B., .7Jl Hubbard, G., 4311
Glassman, J. K., 500 H~D.,gon
Glauber, J. R., 433 Humphrey, D., 44D
Goldman, I., 33411 Hunter, B., 216n
Golino, c., 127 HutchiDa, R. M., 36ft, 35011
Gomperz, H, III Huxley, T. H, log
Good, I. J., 31011 HuypDl, c.,205, 249
Gouldner, A. W., 383ft
Gower, H, 379ft Ibsen, K J., 42S
Gramsci, A., 388n Ihde, A, 12SDt 27011
_ I., 13SDt 223n, 23011
Gray. J. S., 310n James, W., 379D
Greenberg, D. S., In, 3,SD-36n, 43ft, 46n, JIIDIDeI', M., ~IID
48n, 5SDt 6on, 129ft, 28gn, 42 1ft Jenkinson, c., 34sn
Grosberg, P., 32311 Jessop, T. E., 22QIl
Gross, E., 290ft Josephson, B. D., 9sn
Gruber, W. K, 3230 JuriD, J., 221
Grilnbaum, A., 213ft
Gunan, c., 294ft Kapitsa, P. L., 32D
Gunther, R. T., 249ft Kaplan, W., 37ft, 43, 52D, 67D, d6n
Kau1Feldt, A, 32211
Hacking, I., 133ft Kelvin, lArd, 13SDt 158D, 1970
Hadamard, J., 136n Kendrew, J. c., 58n, ~78D, 28.tn, daD
Haptrom, W.O., 1ft Kepler, J., 111ft, 127ft, 188, 239, 349t
Haldane, E. S., 2170 411 ,433
Hall, A Re, 98n, 249ft Kerr, c., 35011
Hall, M. B., 249ft Keynes, J. M., 3540
Hamilton, M., 7SDt 91n, 1030 Killian, J. R., 28311
Hammersley, J. M., 230ft Klein, F., 228
Hankins, T. L., 219ft Knapp, M., xxviii
Hannaway, 0., ~26n Koestler, A, 294ft
Hanson, N. R., 21211 Kohler, W., 1600
Hare, J. K., 114ft Kolbe, A W. K, 26gB
Barnwell, G. P., 65ft Kopnin, P. W., 74ft
Harvey, W., 433 Koyr6, A, 12411
Heath, D. D., 6311 Kronik, D. A., ~501l
Hedgpath, J. W., 349ft Kulm, T., 73ft, 94ft, 116n, 13 1ft, IS8n,
Helmholtz, K von, 38, 39, 40, 8m, 94ft, 261, 2640, 3680
9SDt 25m, 286n, 309Dt 379Dt 388, 414
Hanpe1, C. G., I,ID Labat, Re, 2380
Henry, T., 39311 Labtos, I., xxvii. 151a. 154D. 264n. 2650
Herapath, W., 2S2Jl Ia Metherie, J. c. de, 116n
Heriva!, J. W., 226D Lang, Sir J. G., 55D
Hersh, S. M., 570 Laplace, P. S., 163ft, 226, 227
Hertz, K, 190ft, 21211 Laudan, L., 171D
Hilbert, D., 154Jl, 228, 229 Laue, L WD, 196n
Himmelfarb, G., 38gB Laurent, A, ~7~1
Holitadter, Re, 38gB Lavoisier, A L., 111ft, IISDt 116, 13SDt
Holden, K., B.tn 186n, ~5o, 39311
Hol1iDl, T. H B., 10sn Lazanfeld, P., 38sn
Holton, G., 1358 Leach, J., 1600
Hooke, R., 249 Lcapmaa, Me, 334ft
440 Index ofNames
Lee, H. D. P., 21]1l Minkin, L, xxviii
Leibniz, G. W., 305, 219, 223 Mitebell, J. H., 2500
Lenin, V. I. V., 294Jl Montail"e, M. de, 34SD
Lenoble, R., 12411 Montgomery, D. c., 371n
Leonardo cia Vinci, 206ft Moody, E., 127ft
Lerner, D., 12211 Moore, P. G., 3100
Lewis, J. R., 31011 More, L. T., 2480
Lewis, R. A., 4180, 426D Morgenstern, 0., 86n, 11m, 1230 , Is8n,
Liebig, J. von, 26gB 379ft
LiIge, F., 260 Morley, E. W., 950
Lipson, H. S., 61n Moecovici, S., 306n
Lobachewski, N. I.,411 Muir, R. A., 43
Luce, A. A., 220D Mumford, L., 3600
Luria, S. L., 56n MlJIInYe, A., 26SD
LJICIIko, T. D., 191 391 0, 392, 393ft MUIIOD, A. E., 393ft
McCormmach, R., 2,SOD
McElbcny, V. K., 32Il Nader, R., 42SD
McGuire, J. E., xxvii, 380n Nagel, E., 1710
Mach, E., 30SDt 210D, 211, 21m, 263 Nash, G. D., 52ft
Machlup, F. 210 Nash, L. K., 2SDt 1470
McKeuzie, A. E. Ee, 68n Needham, J., 200, 394D, 43411
McKeon, K., 1710 Newlands, J. A. R., 2520
McKie, D., 263D Newton, Sir I., 98, 107ft, 163n, 192 , 205,
Madane, S., 2880 221, 2D, 248, 249t 253, 308n, 372
MacLaurin, c., 222 Nicholson, M., 42SD
Magini, G. A., 266D ~leberg, H. L, 54Jl, 328, 3Wlt 420D
M.nn"eim, K., 104D Nisbett, R., 344Jl, 3970
Marcuse, H, 434- NordeDski01d, E., 2260
Marquis, D. G., 32 ]1l Norman, D., 3240
Man, K., 21, 312 Novick, S., 349ft
Maxwell, G., 213ft
Maxwell, J. c., IOsn, logo, 1970
May, K. 0., 2730 Oakeshott, M., 104ft
Mayerbo~ H. A., 500 Oldenberg, H., 249
Medawar, Sir P., 41SD Oldham, C. H. G., 3940
Meclvedev, Z. A., 393ft Olson, M., 296n
Meehl, P. E., 3700 Ore, 0., 2480
Mellor, J. W., 1110
Mendel, G. J., 411
Mendelcef, D. I., 252 0, 255 Pacey, A. J., 32211
Menger, K., 21]1l Pagel, W., 433Dt 43SD
Menenne, M., 249 Paneth, F. A., Illn
Merton, R. K., 25SDt 312 Pumekoek, A., 1630, 19SDt 2380
Men, J. T., 168n PII'ICC1suI, 433, 434-
Meyer, V., 25SD Parry, N., xxvii, 42n
Michels, R., 334D Panoas, T., 383ft
Michelson, A. A., 9SD Pasteur, L., 125
Miller, D. c., 9SD Pearson, K., 18, J46n
Miller, G. A., 37011 Ped~ 0., 2660
MiIIibn, R. A., IOSD Peel, Sir Re, 3760
MiUs, C. W., 10, 99Dt 3800, 383ft, 38SD Peirce, c. S., 214
MiUs, D. L., 300ft Pen:iYl1, T., 30211
Miner, H., 560 Peters, R. S., 104ft
 Index ofNames 441
Petit, A. T., 227 Saint Pierre, B. de, 3930
Phillips, J. P., 26gn Sale, R. D., 84D
Pippard, A. B., 2880, 4020 Sakharov, A. D., .llft
Pirani, F. A. E., 9SJ1, 1470 Salviati, F., 3910
Plato,2S, 117, 240 Santillana, G. de, 266~
Pledge, H. T., loon Scheele, C. W., Ilsn
Pogendorf, J. c., 251n Scheler, M., 387ft
Poincare, 8., 1540 Schiller, F, C. S.,73D
Poissoo, S. D., 190, 227 Schi1pp, P., 66ft
Polanyi, M., xxvii,4Sn, 7Sn.14On, 1470. Schimank, 8., 25 1
15 10, 311 , 312, 3810 Scott, R., .3340
PoUock, L. W., 3100 Se1mes, C., 28ft
Poly-, G., 790 Shackle, G. S., 67n
Pompidou, G., 3490 Shapacie, D., 264D
Popowitseh, M. W., 740 Shapiro, B. J., 194ft,
Popper, K. R., 151n, 1910, 2630, 277D, Shapiro, J., 56ft
311 , 312 Sbarikow, J S., 740
Prenant, M., 393D Shaw, G. B., 42711
Price, D. J. de S., 3m , 49ft, 9SDt 3240 Shea, K. P., 349D
Priestley, J., II sn Shryock, R. H., 302n, :M7Dt 3600
Prior, O. H., 387ft Shubik, M., 36gB, 37911, 399n
Proxmire, W., 55D Sidgwick, N. V., 186n
Ptolemy, 137D, 238, 239, 433 Silver, S., 28n
Singer,C.,73D
Rabi, I. I., 24Jl Singer, S. F., 28m
Ramus, P., 248 Singletoo, C. S., 4340
RattaDsi, P. M., xxvii, 1940. 3080. 392n Skolniko~ E. B., 3sn
Ravetz, A., xxviii, 34Sn. 3610 Smeatoo, J., 322
Ravetz, J. R., 387D, 3990, 435n Smith, L., 353n
Ravetz, R. S.,35m Snow, C. P., 24t 211
Reingold, 3sn Solmsen, F., 11711
Richter, I. A., 2060 Spedding,J.630
Roberts, E. B., 323n Spinrad, B. I., 328n
Robertson, J. M., l68n Spohn, W. G., 229ft
Robins, B., 2190 Spnmsen, J. W. van, 255ft
Robinsoo, A. H., 840 Stahl, W., 206n
Robinson, E., 393n Staude, J. R., 38711
Ro~ W. S., 96r Stead, F. M., 3600
Rogers, A. F., 196n Steele, R., 3990
Rokbn, S., 1960 Steffens, L., xxv,96n, 3S9n, 4180. 4190
RoUer, D., 10sn Stein, M., 3790
Roochi, V., III, 304Jl, 207Jl, 2120, 266n, Steinhart, J. S., 1140
3 2 6n Storer, N. t 247ft
Rose, H., 44ft Straehey, c., 91n
Ross, Sir D., 3190 StraDatban,. J. D., 8sn
Ross, S., 3720 SUsskind, C., 137ft
Ross, G. R. T., 217D Swenson, L. S., 950
Roszak, T., 395n Szabadavary, F., 3100
Rothschuh, K. E., loon Szilard, L., :u, 35, 'l.84n
R.oyse, A. B., 2TJD
Rudd, E., 101 Tartaglia, N. t 2480
Rumford, Ctnmt, 2030, 262, 2630 Tatoo, R., 238n
Russell, B., 1540 Thomas, M. c., 368
Rutherford, Lord, 32 Tidman, D. A., 3710
+p ltulez of Names
ToulmiD, S.t xxvi, 171n Westby, G., 3780
Trevor-1loper, H. Re, 86n, 194ft, 3430 WhaIcn, R. W., 28711
Truesdell, C. xxvii, 172n, 193n, 387n Wharton, F. D., 427
Turnbull, K W., 249ft Wheatstone, c., 4311
Tycho BraIle, 433 White, A. D., 18n, 390ft
White, L., 43411
Urquhart, D. J., 49ft Whiteside, D. T., xxvii, 253n
Wiener, N., 33
Weiner, P., 1370
van't Ho~ J. K, 25m, 26gB Wiesner, J. B., 244
Velikovsky, E., 38gn Wi1Iiams, T. 1.,3600
Viddach, A., 379D Wilson, E. B., 7711
Vollmer, H. M., 3000 Wtlson, M. J., 3840
Volta, A., 39 Wmdler, S. C. H., 26gB
WIl Neumann, J., 154Jl, 329
Wittkower, Re, 373ft
Vucinich, A., 411ft Wo~A., 263D
Woo~ A. D., 348n
Waals, J. D., van der, g8 Woo~ K, 158n, 283ft
Waddington, C. H., 3930 Wootton, B"'tmess B., 35 10, 36gB
Walsh, L., 353ft Wootton, J·t 19SDt 278n, 379Jl, 402ft
Wakefield, J. A., 4dn Wundt, W. M., g6n, 379n
Waterston, Re, 252ft Wussing, H., 228ft
WItIOD, E. c., IOSD Wyatt, H. V., 4711
Watson, F. G., 600
WatlOD, J. D., 28gft Yates, F. A., 2TJD, 43411
Watt, J.t 393ft Yuua, M., 6,n
Weber, M., 194ft, 31 2D, 350D, 3870
Webster, c., xxvii. 188n, 34711, 4360 Ziman, J., 44Jl, 730, 26,n
Weiestrass, K. T., 154Jl Z1oczower, A., 67n, 286n
Weinberg, A. M., 16m, 35SD Znaniecki, F., 368n
Wenky, P. G., 391ft Zuckerman, Sir S., 19n, 354Jl
 INDEX OF TOPICS
Here are listed definitions and extended discussions of basic concepts developed in this
work. Compiled with the assistance of Joseph Ravetz.
Academic science, 37-44
Adequacy, 148-52; judgements and at-
teria, 152 -6; aiteria of L and maturing
of a discipline, 156-9
Aphorism, 375-7
Argument in scientific problem, 119-21
'Art', 330, 372-3
Artificiality ofobjects ofscientific inquiry,
110-14
Choice, problem of, 265-8
Citations, etiquette o~ 25~
Cich6-lcience, 385
Conclusion in scientific problem, 125-8
Controlling judgements, 148
Corruption, 418-23
Cnft: scientific work U, 75-6; tech-
niques as c. skills, 101-4; methods as
c. knowledge, 173-6·
Craftsman, 140; craftsman's cosmos, 144-
145
Critical science, 424-31
Cycles: of growth and decline in science,
6,-8; of schools, 226-33; of investiga-
ting technical problems, 323-7; of
investigating practical problems, 339-
346
Data, 76-83; soundness o( 81-3
Decay of information and entropy in-
crease, 202-8
Decisions, between strategies, 265-8;
social aspects of d., 268-71
Descendant problems, lattice of, 191-4
DifFerentiation ofstandardized facts, 209-
211
Dirty science, 27-8; 57
Discovery, compleDty of, 114-16
Engineering, 329-3°
Entrepreneurial science, 46
Entropy increase and decay of informa-
tion, 202-8
Erron in science, 263
Ethics in science, 289-313; final causes
290-4; self-interest, 294-9; DWJage-
ment of property, 'professional eti-
quette', 299-304; professional ethics
and their enforcement, 3~4; levels of
e. commitment, 304-7; social CODtext,
307-11; old and new, 311-13
Etiquette: of citations, 25~; profes-
sional, 299-304
Evidence, 121-5; strength and fit of, 121
Evolution: offacts, 191-4; oftools, 194~;
of objects of inquiry, 202-8
Existing information, 87-8
Experts, 348
Explanation, of. tool, 301
Extension: of tools, 196; of facts, 199-200
Fact, 181-208; introduction, 184-7;
compantive analysis, 187-8; defining
property, 18IHJ1 ; evolution, 191-4;
extension, 199-200; standardization,
199-202; differentiation ofstandardized
fact, 209-11
Final causes of task, 290-4
Folk-lCience, 386-9; and ideological con-
fliet,389-97
FoundatioDs of scientific knowledge; 0b-
scurities, 213-1 3; fallacies of teaching,
21 7-18
Function, and perf011DlllCC ~ 391
Goal, and fulfilment of, 391
'History',373
Hypertrophy, 384
444 Intltz ofTopics
Ideology, 40; in pnctical problems, ]41
Ideological: conflicts, 389-97; sensitivity,
20-1
Immature and ineffective fields, 364-402;
absence of facts, 367-'11; pretence of
maturity, 377-80; teaching of i. disci-
plines, 380-3; in practical problems
397-401
Industriali2ed sc:ience: social problems,
31-68; industrialization of production,
44-7; adulteration of research, 47-5°;
penetration of science by industry, 51-
53; runaway technology, 53-7; monte,
5~; monIity, fn-'1; prospects, 67-8
Information, 83-8; existing i., 87-8;
standard i. .. tools, 89
Intellectually constructed, classes of
things and events, 110-14-
InterpenoDa1 channel of communication,
176-80
Invariance of facts, ISg
Invention, 329-30
Manpower, scientific, 29, 60
Maturity of a discipline, 1 s6-9
MetbocIs: method, methods, methodol-
ogy, 197; generalizations of methods,
167-73; body ofmethods, 172; methods
.. craft knowledge, 173-6; two chan-
nels of communication, 176-80; in-
fluence of methods of scientific know-
ledge, 223-33
MonJe, 57-63
Morality, monls, fnr-7
Motives, 391
Myopic CIlIinccring, 53
Neutrality of science, 415-18
Novelty, manapment o~ 260-,2
Objects of inquiry, 109; evolution and
entropy-inaease, 202-8
Obscurities at foundations of scientific
knowledp, 212-13
Operating cbancteristics of a device,
324-5
Pandoxa: of adequacy, 149, 155; of
value, 166; of obscurity at foundations,
213; of scientific knowledp, 23~
'Philanthropic'science, 434-6
'Philosophy', 372
Pitfalls, 94-101 ; and cycles of practical
problems, 339-46
Practical problems, 33~3; cycle, and
pitfalls, 339-46; scientific aspects,
346-54; comparison to technical prob-
lems, 354-6; practical aspects of tech-
nical problems, 356-63
Pnctical project, 339
Problem-situation, scientific, 135-7
Professional: management of property
and p. etiquette, 299-302; p. ethics and
their enforcement, 302-4
Property, protection of intellectual, 245-
259; earlier social mechanisms, 247-5°;
journals, 250-4; operation of system,
254-6; etiquette of citations, 25~;
industrialized science, 258-g
Pseudo-science, 364-5
Purpose, and achievement of, 291
Q!JaIity conttol in science, 273-88; assess-
ment, 274-8; comparative examples,
278-80; hierarchy of quality conttol,
280-3; leadership and prestige, 283-6;
industrialized science, 286-8
Q!aality conttol of skilled tasks, 294-9
Reckless science, 56
Re-cycling in technical problems, 327-8
Relevance and reliability of information,
84-5
Research report, 182-4; as property, 245-
247
Revolution, scientific, 261-5
'R.omantic' science, 392-3, 43"-5
Runaway technology, 53-7
School, cycle of, 226-33
Science as craft work: data, 76-83;
information, 83-8; tools, 88-94; pit-
faUs, 94-101 ; techniques, 101-4; style,
104-8
Scientific inquiry, analysis by Aristotelian
causes, 116-19
Scientific knowledge: special character,
209-40; obscurity at foundations, 212-
23; historical character, 233-6; unit of,
234; pandoxes, 23~; limits, 237-40
Scientific problem: inteDectual constructs,
110-14; analysis by Aristotelian causes,
116-19; argument, 119-21; evidence,
121-5; conclusion, 115-18; work on,
128-32; definition, 132-5; origins,
135-40
Scientist, definition by contrast to crafts-
man and technician, 143

Shoddy science, 49
Significance, of £acts, 188
Sorcery,33
Soundness, of data, 81-3
Stability, of facts, 188
Standardization: of tools, 197; of facts,
199-203; differentiation of standardized
facts, 209-11
Strategy: of. school, 223-33; problem of
choice between strategies, 265-8
Strength and fit, of evidence, 121
Style, in scientific work, 104-8
Tacit knowledge, in tedmiques, 101-4
Task, aualysis of cfiDal causes', 290-4
Teaching of scienCe: improvements, 207-
208; fallada of 'foundatiODl' teaching,
31 7-18
Tcclmical problem, 331-38; cycle of
investigation, 323-7; comparison to
scientific problem, 337-31; social
upeers, 33 1-7; varieties, 337-8
Intltz o[Topics 44S
Technical project, ~7
Tcclmician, 142
Tcclmiques, 101-4
Technocratic conception of science, 20-3
Tools, 88-94; standard information as
tools, 8g; tookubjcct, 8g; Janguap-
too1, 8g; tool-apcrt, 90; tool-providing
field, 90; asymmetry of provision and
use, 93-4; evolution of tools, 194~;
refinements, 195; atensioas, 196,
_ 197; cxplaDatioDs of
tools, 301
Truth, in science, 17-20
Vilification of. field, ~-6
Vacuous results, 276-7
Value: aiteria and judgements, 15~
objective components, internal and
extema1, 162-5; pcncma1 components,
165-6; problems or value in social
activity of science, 166-,
 GENERAL INDEX

Compiled with the assistance of Joseph Ravetz.

Absolute rigour, 152 Caloric, 127

Absolute space, 263 Capital-intensive research, 44-
Abstract mathematics, 228-3 1 Centres of excellence, 288
Academic cliques, 380 Centre of scientific activity, 67
Academic marketplace, 177 'Change', 213
Aa:eleration, 211 Chemical substance, I I I
Acid, 115 Chemistry, 38, ~
Aeroplane as technical problem, 324 Chlorine, 135
Aggregation of data, 84 Claim-staking,258
'Aim', 292 Clarity of mathematics, 213
Alchemy, 247 'Classic'design, 336
Allison effect, 95 Oub, 177
AItamira cave art, 360 Colonial science, 283
America, 116 Common-tIeuse, artificiality of, 354
Anagram, 248-9 Concave mirrors, 204
Analogical inferences, 120 Conduction of heat in solids, 135
Animated nature, 433 Confirmatory inferences, 120
Answerability,343 Ccmservation of mass, 186
Applied economics, 381 Continental drift, 267
Architecture, 248, 373; pitfalls, 100 Convex lens, 204
Art: imitates Nature, 117; of navigation, Cnft analogy, 117
373; the Arts, 23 Creative generation, 67
Astrology, 238 Criticism in science, 277
Astronomy, 43, 195, 238 Crucial experiment, 263
Atomism, 117 Crucial pheDomeoon, 137
Autonomy of science, 19, 43 Crystallography, 196
Cultural revolution, 394
Curve of quickest descent, 248
Baroque, in mathematics, 229 Cycloid, 249
BibliopaphY,79
Biographies of scientists, 262 Dam, functions o~ 358
Biology, ~ Darwinism, 389
Biological species, 112 DDT, 357
Blunders, 47 Deductive inferences, 120
Bodega Head, 349 Dehumanization and disenchantment, 64
Bonds, 86 'Democracy' of science, 394
Boondoggling, 54, 328 Detoun in advance of science, 224
Bounty,52 Diagnostic problem, 103
Bow and arrow pandox, 206 Di1rerential calculus, pitfalls, 219
Boyle'. Law, 204 Direct-square law of association, 101
Bulgarian pip, 86
Bureaucracy, ~ Early diwmination scheme, 49
 General Intlez 447
Ecology, u6 Historical research, 139
Economics, 379 'History', 174
Einsteinian revolution, 263 History of art, 104
Electric current, 214 History, philosophy c( I~
Electricity,32 3 L'HOtIIIIIe 1IIM_ne, 158
Electricity meter, tapping o~ 2C)6 Housing and the poor, :US
Electrodynamics of moving bodies, 135 Humanistic scholarship, 247
Electrolytic dissociation, 151 Hypothesis, 187
Electronic computers, 91 Hypothetico-deduetive method, 133
Element, III
Embalming science, 350 Image, 213
Encyclopedias, 206 Inclined-plane experiment, 1~3
Energy,113 Inductive inferences, 120
Engineering, 248 Inert gases, 186
Environment, 27
Errors in science, 263
Esoteric character of science,I.
Estimation of 'significance', 84
'Ethos' of science, 31~
'Inferior stocks', 346
Infinite degrees of slowness, 3~9
Information aisis, 47
Initial insight, 138
Intellectual genealogy, 105
Eugenics, 391 InteUectual honesty, 311
Experience, in Gali1eo, 124 Inte11ectuals, 427
Experimental psychology and ethics, 96-7 InteDec:tua1ity, 355
Exploitation of scientific results, 255-6 Inte1ligencu, 249
Interferometer, 95
F==fIIII, 212 Invalid patterns of argument, 153
Factory inspectorate, 348 Invective, 269
Falsification, 151 Inverse-square Jaw of gravity, 253
'Fiddle', 2CJ7 Iron law of oliprchy, 344
Field, 113
Folk-history of science, 263-3 Know1edp-industry, 21
Folk-lCiences of practitioners, 351
Foo1-proof models, 197 Latent function, 1 95
Formally-valid patterns of argument, 120 Law, 187, :u6.
Foundatious, 154- Law of hDel, 310
Fourth Law of Thamodyuamics, 76 Lift-pumps, 137
Fnuds, scientific, 36 Literary aiticism, 1 74
French physical science, early 19th Logical empiricism, 171
century, 226-8
Magic, 63, 144, 1 74
Gaps in scientific creatinty, 68 Marxism, 19, 312, 388
Gam-warfare research, 65 Mass, 113, 311
German universities, 286 Master-drain, 360
GIGO, 370 Mathematical culture, lack of, 1 58
Goal-diaplacement, 341 Mathematics, 248, 288; nineteenth-
'Grand problem', 163, 1M century, 199
Grantsmanship, 45 Measunment, principles ~ ~14
Gresham's Law for inteDectual property, Mechanical philosophy, 64
281 Mechanics, 801II'CeI in aperience, 205
Medicine, ~, 346; nineteenth-century,
Hazards of sponsored research, 51 302
Heliocenttic system, 127 Mediocre raean:h, 288.
Hierarchy of conttol, in science, 281 Mediocrity, competence ~ 123
Hippie, ~7 Mesmaism, 367
Hiroshima and Napaki, 64 Metaphysics and dopDa, 18
448 General Intlez
MicrOICOpe, 197 Physics, 60, 61, 263
Military tedmology, 57 Planned science, 45
Mill'. canons of induction, 153 Plasma physics, 371
Miaion-oriented raean:h, 53 Poetry, 175
Model, 187 Polycenttic structure, 270
Molecular biology, 288 'Positive', 387
_ , 1 51 Pnctical men of affairs, 345
Moon, shadows on, 73 Precunon of Gali1eo, 127
Multivenity, d6 Price, 113
Mysteries of the calculus, uo Priority~255
Mystical cxpcricnce, 175 Probabilistic inferences, 120
ProbIeDHhiCt, 264
Nagasaki, 64-5 ProcIamaticms of method, 170
National Animal Speech Agency, 55 Production of work by heat-engine, J 35
National styIcs of scieDc:e, 35 1 Profeaionalization, 306
Natural magic, 13 Profit, 54t 113
Natural theology, 308 Propas, 30
Nature's abhorrence of. vacuum, 137 Pseudq-precision, 158
NIIIWIJalo.p,", 94 Pseudo-property, 287
NepocI, 353 Psychology, 174, 373~ 37~
Newton-Leibniz dispute, 249 Psychology of invention, 136
~IIPI", 378 Public c:baDnel of communication, 176-7
Nobody, rule by, 363 Pablieation-poin1l, 50, 287
Non-book, 48 Punch-eards, 352-3
NOD-UDiform motion, 98 Pure science, 431 164
Normal 8CieDee, 13 1 PuzzIc.eolviDl, 131
N-nya, 95
Nuclear strategy, 354; as gematria, 399 Q!Jantity of heat, 203
Nuclear wapcms, 27
Number, 214 Racket, 395
RantHinder, 326
'Objective', 393 RatioDal mechanics, 172
Obscurities in mc:cbanics, 219 RatioDaIity, 144, 387
0fIic:ial denial, 349 Reception of electrical signals, 137
0fIic:ial patronage of science, 168 Recruitment to science, 5~
Operationalism, 202 Ilefincment of tools, 195
Optics, III Refutation of hypotheses, 172
0rpDizcd common sense, log IlcCuted &da, 186
Oudyins data, 79 RepIatoI"y commissions, Hfo-cyc1e of, 359
Oxypn, I I I R.csearch, 13-1], 15-16
Ozniw, liS, 135 Research assistants, 96
Oxy-muriatic bile, 135 Raidues of epochs, 209
Responsibility of scientists, 19, 25
Pandipn, 264 Responsibility without power, 423
Paradise lost, 33 Rumford myth, 263
Parody,36g
Particle acceIeraton, 135 Sanitary movement, . -
Pardy-babcl idea, 310 'Scholastic' method, 247
'Perfect' problcm, 193 Science policy, II, 36
Period of simple pendulum, 138 'Sc:ienccI', 387
PetroI-eDgine, 333 Scientific community, 243
PhiIoIophy of technology, 317 Sc:ieDtific ethic, 30g
PbJoPston, 127, 262 Scientific manpower, 21
Physical CODSWlts, IU Scientific method, 17, 147
 General Indez 449
Scientific revolution, 17, 387 Tartrates, crystaI1ine forms, 125
'Scientist', 372 Teachinl, 59
'See also', 273 TccImical intellipntlia, 428
Senescence of discipline, 402 Technological fix, 355
Shoddy technology, 334-5 Technological forecasting, 369
Significant digits, 216 Technology of me home, 361
Skills of experimental research, 82 Telegraph, 39
Skills and knowledge, 217 Telescope, Galileo's, 266
Slavery, 426 Temperature, 210
Slide-rule, 197 Testing hypotheses, 129
Sober drunkenness, 137 Theology, 174
Socialism, 294 Theory, 187
Sociology, American, 368 'True Boiling Point', 324
Sodium Hydroxide, 96
Sorcery, 65 Uniform circuJar motion, 186
Spaco-race, 57 Universities, 39
Spurious confirmations, 370 Urban motonrays, 3M
Standard deviation, 203 Utility,39
Statics, 205
Statistical s a _ e , 158
Statistics as folk-sclence, 396 Valency, 204-
Statistics, national, 123 Values of fundamental physical constants,
Statistics, vacuous, 277 85
Student, 14-15 Variable, 213
Supernatural explanations, 12 Ym'lII '""IOri,
fiill, 149
Vitalism, 127,263
Supersonic airliners, 23, 54
Swing from science, 27
Synthetic dyestuffs, 323 Warfare of science with theology, 18, 390
Systm.e l"terut;tnUll, 326 W;IInI#Uft, d6

Taoism, 434 Zeno's paradoxes, 213

 8oclolo~
PhlIoeophy

SCIENTD'IC KNOWLEDGE
AND ITS SOCIAL PROBLEMS
Jerome R. Ravetz
With a new Introduction by the author
Science Is continually confronted by new and d1fDcult social and ethical
problems. Some of these problems have arisen from the transformation of
the academic sclence of the prewar period into the Industrialized science of
the present. Traditional theories of science are now widely recognized as
obsolete. In ScI.entfIIc Knowledge and Its SocIal Problems (or1g1nally published
in 1971), Jerome R. Ravetz ana1yzes the work of science as the creation and
investigation of problems. He demonstrates the role of choice and value
judgment, and the inevitabWty of error, in sclentlflc research. Ravetz·s new
introductory essay Is a masterful statement of how our understanding of
sclence has evolved over the last two decades.
-SctentYfc Knowledge and Its SocIal Problems Is both monumental and remark-
able ... a major contribution to the understanding of sclence.- -Jonathan
Rosenhead, New Sctenttst and Sc1ence Journal
-mhoughtful, incisive, scholarly, lucid, humane and sane. Everyone-but
everyone-within or concerned about science should read It slowly and
carefully.- -John ZIman, Nature
-nus Is a penetrating and novel account of the sclentlflc activity. - - Ttmes
Uterary Supplement
-Scten.tf/lc Knowledge and Its Soc1.al Problems, Jerome R. Ravetz's major
contribution to the history and philosophy of sclence, should be of interest to
anyone concerned with change in education.- - Robert McClintock. Teachers
College Record
-Scten.tf/lc Knowledge and Its SocIal Problems deserves to stand for some time
to come as one of the landmarks of the new cI1t1cal spll1t which Is transfonnJng
science from within in ways whose political and cultural bnportance cannot
be exaggerated.- -Theodore Roszak. The Nation
About the AIIthor

Jerome R. Ravetz taught history and

phIlosophy of science at the University of
Leeds; he Is now based In London, as a
consultant and independent scholar.

Ubrary of Congress: 95-32177

PrInted in the U.S.A
Cover design by Karen M Surowiec

ISBN 1-56000-851-2
90000>

9 781560008514 I ISBN: 1-56000-851-2 I